2014-08-01 06:32:24 Hi ! 2014-08-01 06:54:26 Hello... I have two routers 'out-there' (remote) with Alpine 2.2.2 running of USB (or CF card) - they use UCARP for HA ... 2014-08-01 06:55:23 Every now and then they start a 'rebooting cycle' - the one, that 's a UCARP 'master' reboots, the other one takes over, the couple of minutes later that other one reboots and the first one takes over and so on... 2014-08-01 06:56:19 It doesn't seem to be linked to temperature as they are both in a chilled room (and 'sensors' shows fairly normal core temperatures)... 2014-08-01 06:57:04 Are there any known bugs with 2.6.38 grsec kernel such that a network 'attack' could cause these reboots ? 2014-08-01 06:59:30 Hmm, like that one, perhaps? http://www.securityfocus.com/archive/1/518014 2014-08-01 07:23:59 2.6.38 has likely a bunch of bugs... 2014-08-01 07:26:15 http://bugs.alpinelinux.org/projects/alpine/issues?utf8=%E2%9C%93&set_filter=1&f[]=status_id&op[status_id]=c&f[]=subject&op[subject]=~&v[subject][]=kernel&f[]=&c[]=project&c[]=tracker&c[]=status&c[]=priority&c[]=subject&c[]=assigned_to&c[]=updated_on&group_by= 2014-08-01 07:44:11 Oh my. :D 2014-08-01 08:10:06 i think i see a new kernel vuln almost every week nowdays... 2014-08-01 08:10:35 and i know atleast 2 vulns not fixed in the latest -stable kernels 2014-08-01 08:17:47 Not good... 2014-08-01 08:19:36 Any idea how to catch what's causing the problem on a remote, headless, 'run-from-ram' Alpine machine? Maybe I could find some work-around / protection... 2014-08-01 08:20:25 i have no idea 2014-08-01 08:25:23 danci1973, enable logging to remote server, or disk 2014-08-01 08:29:21 fabled: I am logging to a remote server via syslog, but there's nothing (apart from a 'gap' in normal log entires)... It seems that whatever happens it doesn't have a chance to send it over the network anymore. 2014-08-01 08:29:41 And the logging server is local to those routers. 2014-08-01 08:30:12 you can get console output? 2014-08-01 08:32:07 fabled: Not sure what you mean? 2014-08-01 08:32:21 fabled: These routers are currently headless and remote to me. 2014-08-01 08:33:12 on many places i have console on ttyS0, and it's accessible via remove serial port server 2014-08-01 08:33:24 though, not all 2014-08-01 08:33:32 I'm trying to get a local guy there to connect a monitor to one and to potentialy photograph the screen output - if something shows up... 2014-08-01 08:33:35 it's hard to debug if there's no logs of what's happening 2014-08-01 08:33:46 fabled: No, unfortunately no serial console there (yet). 2014-08-01 08:34:16 is the same issue happening on all places, or just few places? 2014-08-01 08:34:29 fabled: Just one place... 2014-08-01 08:34:37 could be hardware failure too then 2014-08-01 08:34:58 fabled: But to be fair, I don't have the exact same configuration (hardware or Alpine version) on all places. 2014-08-01 08:36:25 Yeah, it could be hardware failure, but there are two routers (same HW, though) showing the same behaviour. They are on an UPS, so AC should be stable (there are other servers on the same UPS that don't show problems). 2014-08-01 09:02:59 OK, I got something now... I was running 'iftop' and when the crash happened, I got a 'segmentation fault' and a couple of 'kernel: PAX ....' lines to my syslog... 2014-08-01 09:04:55 Could this be RAM / memory issues? 2014-08-01 09:05:44 could be that your are running out of memory 2014-08-01 09:05:49 it could be memory is corrupt 2014-08-01 09:06:14 depends on what more than 'segfault'2 and 'kernel: PAX...' 2014-08-01 09:07:35 http://paste.ubuntu.com/7922208/ 2014-08-01 09:08:22 The 'xxx.xxx.xxx.xxx' is my outbound IP ... 2014-08-01 09:09:27 I don't think I'm running out of memory - both have 1GB... But they do have DIMMs from the same batch, so memory fault is an option... 2014-08-01 09:09:41 that is iftop that is segfault 2014-08-01 09:09:53 and it segfaults due to its doing bad things 2014-08-01 09:10:03 and pax beats it down with a hammer 2014-08-01 09:10:12 that could even be a bug in iftop 2014-08-01 09:10:18 and might not be related your other issues 2014-08-01 09:11:51 I find it a bit strange that these lines appeared as the router crashed. 2014-08-01 09:12:11 <^7heo_> hey people 2014-08-01 09:12:22 <^7heo_> I am trying to install alpine on an intel NUC 2014-08-01 09:12:33 <^7heo_> and the USB keyboard isn't recognized. 2014-08-01 09:12:39 <^7heo_> (i.e. I can't even login as root) 2014-08-01 09:12:46 <^7heo_> any idea why/how to fix it? 2014-08-01 09:12:53 <^7heo_> ncopa: maybe you know? :P 2014-08-01 09:39:08 thats weird 2014-08-01 09:39:34 ^7heo: does the key work if you boot other distro? 2014-08-01 13:02:48 ncopa: I finally got some 'kernel' messages logged in my syslog - can you please have a look? http://paste.ubuntu.com/7923820/ 2014-08-01 13:17:15 looks clearly kernel bug 2014-08-01 13:19:56 ^7heo - try changing USB Legacy value in BIOS? 2014-08-01 13:21:37 danci1973, is it smp machine? 2014-08-01 13:22:13 seems snmpd triggered the crash, sounds like specific type of SNMP GET that triggers 2014-08-01 13:22:15 it 2014-08-01 13:22:25 fabled: It has two cores... 2014-08-01 13:24:57 danci1973, disable snmp? or upgrade? 2014-08-01 13:25:58 fabled: I'll try disabling it, of course... 2014-08-01 13:26:44 CPU is Atom D510, 2 cores, 4 threads, ... 2014-08-01 13:27:38 Although I have my doubts - snmp really shouldn't be able to crash the kernel... 2014-08-01 13:35:27 danci1973, yes, it's kernel bug 2014-08-01 13:35:43 seems like snmpd receives message, and reads info from kernel via netlink and that crashes it somehow 2014-08-01 13:35:45 fabled: Unfortunately, it doesn't seem to be the case... 2014-08-01 13:35:49 hmm 2014-08-01 13:35:50 ok 2014-08-01 13:35:54 fabled: I stopped snmpd and it happened again. 2014-08-01 13:36:09 perhaps run memtest? 2014-08-01 13:36:54 fabled: That would be great, but it's a bit hard since the machines are remote and headless... 2014-08-01 13:37:50 fabled: Furthere more - there are two machines - it would be a collosall coincidence if both DIMMs failed at the same time... 2014-08-01 13:38:39 OK, no1 just got back online - it's the first time booted withot starting snmpd ... 2014-08-01 13:38:59 sounds weird issue 2014-08-01 13:39:03 and kernel bug 2014-08-01 13:39:13 recommend upgrading SW 2014-08-01 13:39:13 Yup... 2014-08-01 13:39:20 it's kinda ancient 2014-08-01 13:39:45 I'm preparing a new 'image', but it won't be easy replacing it on the running machines that keep rebooting... 2014-08-01 13:40:38 it just started suddenly to do that? 2014-08-01 13:41:35 In that order of magnitude - yes. Before it was just once every few weeks ... 2014-08-01 13:42:07 Besides - I'm not 100% sure it wasn't something else before ... 2014-08-01 13:43:42 The only change I can see is increased network traffic compared to 'normal' days... 2014-08-01 13:44:00 (which is what I use snmpd for) :) 2014-08-01 13:52:11 well there was a fairly recent UDP issue foudn in kernel 2014-08-01 13:52:22 could be bad guys trying to exploit it 2014-08-01 13:53:20 ncopa: Do you remember a bit more? So I can find if there's some sort of protection workaround possible? 2014-08-01 13:55:26 http://www.cvedetails.com/cve/CVE-2013-4470/ 2014-08-01 13:55:53 Thanks. 2014-08-01 13:56:37 danci1973, if you suspect it's network traffic related, you might want to disable all NIC offloads 2014-08-01 13:57:23 fabled: Is this something that's enabled by default? Cause I didn't enable anything like that... 2014-08-01 13:57:31 fabled: Or in another words - how? :D 2014-08-01 13:57:34 danci1973, depends on hardware 2014-08-01 13:57:37 ethtool 2014-08-01 13:57:56 ethtool -k ethX -- to inquire 2014-08-01 13:58:07 and ethtool -K ethX args... -- to modify 2014-08-01 13:58:25 There are 6 Intel gigabit interfaces - one 82567V-3 and five 82583V ... 2014-08-01 13:59:27 rx-checksumming, tx-checksumming, scatter-gather, tcp segmentation offload, generic segmentation offload are all ON 2014-08-01 13:59:39 udp fragmentation offload is OFF 2014-08-01 14:02:13 i'd probably try turning off tcp seg. and generic seg. offloads first 2014-08-01 14:02:25 In '/etc/network/interfaces', can I have severl 'up' lines for an interface ? 2014-08-01 14:02:34 Or do I have to format it all in one? 2014-08-01 14:03:31 you can have several up lines 2014-08-01 14:04:56 Thanks... 2014-08-01 14:05:13 I disabled all of those and hav my fingers crossed. 2014-08-01 14:08:26 Thank you for all your help - I have to go now... 2014-08-01 14:28:10 Hey... 2014-08-01 14:28:52 Maybe the NIC offloading thing is on the right track - it's 28 minuts without a reboot... 2014-08-02 08:42:22 apk upgrade -v 2014-08-02 08:42:27 1 errors; 418 packages, 2287 dirs, 36101 files, 1168 MiB 2014-08-02 08:43:01 ^ gives me just that and seems like update/upgrade went fine, it there a way to dig into it more? 2014-08-02 09:35:32 <^7heo> ncopa: the keyboard works with slackware I believe. 2014-08-02 09:35:38 <^7heo> ncopa: I'll check on monday :) 2014-08-02 09:53:50 hi ! 2014-08-02 10:42:21 melville: apk fix maybe 2014-08-02 10:55:00 can I ask mdev to re-read disk ? I just partitionned my drive, but I can't see either sda1 or sda3 2014-08-02 10:55:06 only sda and sda2 2014-08-02 10:55:40 lsblk shows the 3 partitions though 2014-08-02 11:18:45 z3bra: you could create the block devices with mknod 2014-08-02 11:19:49 i think partprobe worked for me 2014-08-02 11:26:38 thanks darkfaded 2014-08-02 11:41:41 melville: one thing that doesn't work with mdev is doing kpartx -a /dev/vm/image 2014-08-02 11:41:50 you won't get the devices in /dev/mapper 2014-08-02 11:42:07 but seems there mknod would be a good fix 2014-08-03 13:26:26 Hi ! 2014-08-03 13:26:39 Does the grsec kernel support xfs by default ? 2014-08-03 13:27:01 because I can't get my server to boot from an XFS partition 2014-08-03 13:27:18 extlinux.conf seems ok (good UUID, xfs as a module) 2014-08-03 13:27:31 I recreated the initramfs, to be sure 2014-08-03 13:28:13 But after the extlinux menu, I end up with a "Can't mount root" error, dropping me in emergency shell 2014-08-03 13:35:54 any verbose log messages? 2014-08-03 13:36:14 maybe grsec deny module auto load 2014-08-03 14:59:23 z3bra: modules=xfs kernel boot param? 2014-08-03 16:00:22 i don't think modules=xfs will help 2014-08-03 16:01:15 you need xfs in initramfs, otherwise you can't mount / partition if it is xfs 2014-08-03 16:01:41 to load module you need mount / first 2014-08-03 16:04:44 z3bra, for / partition on xfs you either need xfs in-kernel or xfs module in initramfs 2014-08-03 16:10:15 i would leave /boot partition alone on ext 2014-08-03 16:26:49 did one of you look into ceph on alpine> 2014-08-03 17:52:58 a long time ago 2014-08-03 17:53:06 it has some glibc dependencies 2014-08-03 18:35:02 kocka: thanks, then i'll probably be best off attaching storage via /sys/bus/rbd/... instead of trying to get rbd to build 2014-08-03 18:42:50 isn't ceph c++ app? 2014-08-03 18:43:43 yes 2014-08-03 18:44:10 i just want rbd devices in dom0, don't care about the rest 2014-08-03 18:44:36 osd (storage) needs to run in domU anyway due to data corruption by tcp offloading with xen 2014-08-03 18:45:04 since noone fixes that... all the c++/glibc stuff would be in vms running whatever :) 2014-08-03 18:46:07 a friend and i made a lot of it already working, just we can't use it on alpine (yet) 2014-08-03 18:48:09 so you need only librbd? 2014-08-03 18:49:07 https://packages.debian.org/sid/librbd1 2014-08-03 18:49:34 no not even that 2014-08-03 18:49:36 weell 2014-08-03 18:49:48 at minimum the rbd kernel module 2014-08-03 18:50:19 librbd and the rbd utility (in ceph-common on debian) would be the version making my life not hell 2014-08-03 18:51:03 if you have the kernel module, you can add storage via /sys/bus, and with the "rbd" utility you can just do "rbd map foo" 2014-08-03 18:52:50 CONFIG_BLK_DEV_RBD=m is enabled by default, if this is right module? 2014-08-03 18:52:56 yes 2014-08-03 18:53:03 it wasn't there when i checked 2014-08-03 18:53:05 or i missed it 2014-08-03 18:53:06 sec 2014-08-03 18:53:35 88 /lib/modules/3.14.15-0-grsec/kernel/drivers/block/rbd.ko 2014-08-03 18:53:36 ah, maybe because i checked at home, i'm on 2.7 there 2014-08-03 18:53:43 using edge here 2014-08-03 18:54:02 yeah i had to roll back because no hvm VM was working any more after going to 3.x 2014-08-03 18:54:25 also not on 3.x 2014-08-03 18:54:40 i'll make a edge test vm later this week 2014-08-03 18:54:44 sounds great 2014-08-03 18:55:30 would be really nice ceph working on alpine 2014-08-03 18:59:33 i could be wrong, but for ceph boost could be potentially more of a problem than musl, since it uses stdc++. if i remember i did read rants about boost breaking api (or was it abi?) between minor versions. 2014-08-03 19:03:27 it's all pretty nightmarish 2014-08-03 19:08:42 "It's packaging other people's software that makes system administrators violent people." - Theo Schlossnagle via https://www.youtube.com/watch?v=ieCTIPG43no (~25 minute story describing experience when one software breaks after another) 2014-08-03 19:09:00 *breaking 2014-08-03 19:09:08 rofl i better won't watch that today 2014-08-03 22:12:41 watched it now :) 2014-08-03 22:12:59 the one thing he has wrong is the turnover for such a tcp offload debugging 2014-08-03 22:13:09 we had such a bug on aix in ~2005 2014-08-03 22:13:17 it took < 6 hours 2014-08-03 22:13:32 and that is, sadly, pretty typical of the devops crowd 2014-08-03 22:14:07 maybe we got the patched kernel a day later, but tracing down the issue and informing people was all fine and totally standard 2014-08-03 22:14:27 the only difference in enterprise IT you don't go all fucking amazed because of a bad hw bug 2014-08-04 08:19:57 Good morning all 2014-08-04 09:18:37 hi 2014-08-04 09:26:31 yo 2014-08-04 12:41:54 any idea why this isnt working: 2014-08-04 12:41:57 xen_commandline : dom0_mem=dom0_mem=2048M tmem tmem_compress tmem_dedup iommu=1 dom0_vcpus_pin cpuidle numa 2014-08-04 12:42:03 daveh0003:~# grep -i total /proc/meminfo 2014-08-04 12:42:03 MemTotal: 38497836 kB 2014-08-04 12:42:48 i don't need 40GB+ ram in my dom0 it just slows down vm creations and creates fragmented memory 2014-08-04 12:45:30 darkfader: dom0_mem=dom0_mem is certainly not valid, don't you want just dom0_mem=2048M? 2014-08-04 12:45:57 ahhh!!!!! 2014-08-04 12:46:00 thanks 2014-08-04 12:46:04 and sorry 2014-08-04 12:46:39 never noticed :) 2014-08-04 15:28:12 heads up to the community: I did edit http://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB#Creating_a_bootable_Alpine_Linux_USB_Stick_from_the_command_line 2014-08-04 15:28:50 <^7heo_> I did add the recommendation to use a virtual machine rather than your own machine, as it is safer and simpler. 2014-08-05 08:00:54 was here some asking about asterisk on 3.0/edge? 2014-08-05 08:01:18 FYI, after yesterdays and todays fixes, I'm now able to run asterisk and at least basic config works 2014-08-05 08:51:45 <^7heo_> moin here 2014-08-05 08:52:09 <^7heo_> ncopa: I'm currently trying to get the usb to work in the NUC I'm booting with the alpine installation iso (64 bit) 2014-08-05 08:52:16 <^7heo_> it doesn't get any usb signal 2014-08-05 09:42:58 ^7heo, I was planning to get a NUC, but read few comments on forum and thought to wait and see more reviews 2014-08-05 09:42:58 https://communities.intel.com/community/tech/nuc 2014-08-05 10:04:12 ^7heo did you try changing "legacy usb" setting in bios 2014-08-05 10:04:35 intel seems to recommed this for nuc on openelec (linux distro) http://www.intel.com/support/motherboards/desktop/sb/CS-033935.htm 2014-08-05 10:07:23 ^7heo_, did you get my message? i'm not sure which account you're using 2014-08-05 10:07:35 <^7heo_> melville: currently I'm using ^7heo_ 2014-08-05 10:07:40 <^7heo_> the other one is @home 2014-08-05 10:07:46 <^7heo_> and I didn't get your message, no. 2014-08-05 10:08:02 did you try changing "legacy usb" setting in bios 2014-08-05 10:08:10 intel seems to recommed this for nuc on openelec (linux distro) http://www.intel.com/support/motherboards/desktop/sb/CS-033935.htm 2014-08-05 10:08:44 <^7heo_> hmm 2014-08-05 10:08:48 <^7heo_> going to check that, thanks. 2014-08-05 10:09:00 <^7heo_> I use too many computers right now, so it's kinda hard for me to remember everythnig :P 2014-08-05 10:09:03 <^7heo_> everything* even 2014-08-05 10:21:22 <^7heo_> damn, I can't get it working... 2014-08-05 10:27:27 <^7heo_> ncopa: CRUX boots fine and gives me the keyboard; but alpine does not. 2014-08-05 10:27:31 <^7heo_> =/ 2014-08-05 10:29:36 <^7heo_> I think I remember that I have tried to install CRUX once. 2014-08-05 10:29:43 <^7heo_> why don't I remember that? 2014-08-05 10:54:21 look into dmesg? can you get this or at least take picture with phone? 2014-08-05 11:09:44 <^7heo_> melville: I can't type anything 2014-08-05 11:09:53 <^7heo_> if I could type, I would start the setup, not look into dmesg 2014-08-05 11:14:16 keyboard isn't working for syslinux also (with and without legacy usb)? 2014-08-05 11:15:15 ^7heo_: have you tried other distro's? 2014-08-05 11:15:53 could be your keyboard needs another module. 2014-08-05 11:16:16 but ive also had issues with usb keyboards before 2014-08-05 11:16:38 <^7heo_> clandmeter: I'm currently using CRUX, works well 2014-08-05 11:17:19 can you lsmod |sprunge? 2014-08-05 11:17:26 or simular 2014-08-05 11:17:28 similar 2014-08-05 11:18:09 <^7heo_> no, no keyboard. 2014-08-05 11:18:19 i guess clandmeter means in crux 2014-08-05 11:18:27 yes 2014-08-05 11:18:27 <^7heo_> ah 2014-08-05 11:18:40 which kernel does crux have? 2014-08-05 11:18:46 <^7heo_> I need to go but I could try 2014-08-05 11:19:32 lsmod | curl -F 'sprunge=<-' http://sprunge.us 2014-08-05 11:19:33 <^7heo_> sprunge command not found. 2014-08-05 11:19:37 <^7heo_> however 2014-08-05 11:19:56 <^7heo_> lsmod | grep sprunge returns nothing 2014-08-05 11:20:05 nah try; cat bin/ching | curl -F 'sprunge=<-' http://sprunge.us 2014-08-05 11:20:08 <^7heo_> yeah 2014-08-05 11:20:11 <^7heo_> okay 2014-08-05 11:20:14 <^7heo_> I don't use google 2014-08-05 11:20:20 <^7heo_> so I didn't notice it was a pastebin 2014-08-05 11:21:09 <^7heo_> curl command not found 2014-08-05 11:21:10 <^7heo_> sorry 2014-08-05 11:21:16 :) 2014-08-05 11:21:26 <^7heo_> anyway 2014-08-05 11:21:27 <^7heo_> gtg 2014-08-05 11:21:27 pastbin is fine 2014-08-05 11:21:34 <^7heo_> I'll paste that asap 2014-08-05 11:21:40 <^7heo_> in 3 hrs or something. 2014-08-05 11:21:48 np if im here 2014-08-05 11:22:41 ^7heo, also check if keyboard is working under syslinux, if it does you could change linux boot parameters to verbose or so. 2014-08-05 12:51:57 <^7heo_> melville: yeah right, but the alpine boot cd doesn't expose syslinux does it? 2014-08-05 13:29:50 ^7heo_, isolinux does provide "boot:", just it start very fast. you can stop it by typing. 2014-08-05 13:30:35 Boot: /boot/grsec initrd=/boot/grsec.gz blah blah blah 2014-08-05 13:30:51 just look at /boot/syslinux/isolinux.cfg on iso 2014-08-05 13:32:03 <^7heo_> for now I try to understand why lilo is giving me a "No boot signature in partition" message. 2014-08-05 13:38:54 did you change ahci in bios by any chance? this could change device names if i remember 2014-08-05 13:43:20 <^7heo_> melville: yeah but there's one disk, it's /dev/sda1 2014-08-05 14:06:19 <^7heo_> melville, clandmeter: okay, now I can paste your stuff. 2014-08-05 14:06:26 <^7heo_> so, you want the output of lsmod on the CRUX? 2014-08-05 14:09:18 clandmeter: can you lsmod |sprunge? 2014-08-05 14:10:09 maybe pinging will bring him 2014-08-05 14:11:13 im here 2014-08-05 14:14:01 check lsmod and see if you see any modules that are keyboard related 2014-08-05 14:14:27 i dont know what CRUX is 2014-08-05 14:14:31 is it linux based? 2014-08-05 14:15:27 <^7heo_> clandmeter: it's a linux yes. 2014-08-05 14:19:09 <^7heo_> I'm getting the network up on the computer and I'm sending that info to you 2014-08-05 14:21:19 <^7heo_> clandmeter, melville: http://sprunge.us/PBhd 2014-08-05 14:22:36 hid_cherry 2014-08-05 14:22:56 could be thats not loaded by default? 2014-08-05 14:23:39 # CONFIG_HID_CHERRY is not set 2014-08-05 14:23:54 <^7heo_> ah yeah it's a set of cherry devices. 2014-08-05 14:23:55 <^7heo_> right. 2014-08-05 14:24:02 <^7heo_> might be why. 2014-08-05 14:24:19 no hid_cherry in defaul kernel on musl 2014-08-05 14:24:27 *default 2014-08-05 14:24:45 <^7heo_> well, then, the kernel of the alpine linux should be updated to include that module 2014-08-05 14:25:04 <^7heo_> because really, it's a very common brand, here (Germany0 2014-08-05 14:25:11 <^7heo_> s/0$/)/g 2014-08-05 14:25:34 http://git.alpinelinux.org/cgit/aports/tree/main/linux-grsec/kernelconfig.x86_64#n4360 2014-08-05 14:26:31 <^7heo_> but why is it avoiding THAT many drivers? 2014-08-05 14:26:39 <^7heo_> I mean, belkin... 2014-08-05 14:26:43 <^7heo_> it's also common 2014-08-05 14:28:59 <^7heo_> SAITEK is not set?!!! 2014-08-05 14:29:15 <^7heo_> that should also be changed, saitek is a very common brand of gaming keyboards... 2014-08-05 14:29:31 <^7heo_> same for steelseries... 2014-08-05 14:29:53 submit a patch/bug 2014-08-05 14:29:55 <^7heo_> they should at least be put in modules. 2014-08-05 14:29:58 <^7heo_> yeah I wil 2014-08-05 14:29:59 <^7heo_> will* 2014-08-05 14:31:52 <^7heo_> clandmeter: where to submit patches? 2014-08-05 14:31:58 <^7heo_> ml I gues? 2014-08-05 14:32:10 <^7heo_> s/s\?$/ss?/g' 2014-08-05 14:33:42 yes 2014-08-05 14:33:47 or github 2014-08-05 14:34:06 github you would be the first :) 2014-08-05 14:35:54 ^7heo_: you could also open a bug report, what you prefer 2014-08-05 14:36:21 <^7heo_> meh, I suppose that I can login into github 2014-08-05 14:36:28 <^7heo_> and submit a pull request 2014-08-05 14:36:42 <^7heo_> is it fine to set the keyboard brands I know to 'm'? 2014-08-05 14:36:48 <^7heo_> (afaik, using modules for them) 2014-08-05 14:37:25 i think so 2014-08-05 14:37:38 <^7heo_> okay 2014-08-05 14:37:46 maybe some of them should also go into intiramfs 2014-08-05 14:37:47 <^7heo_> I write the doc I have to, and then I do that. 2014-08-05 14:38:10 keyboard is the least we could support ;-) 2014-08-05 14:38:37 but it does sound strange we need so many modules for just keyboards. 2014-08-05 14:40:00 <^7heo_> meh, I suppose that different constructors have different usb controllers and firmwares. 2014-08-05 15:32:16 hi 2014-08-05 15:34:10 trivia: Is there way to increase dom-0 memory at runtime without reboot? 2014-08-05 15:34:39 xl mem-set/max Domain-0 can certainly decrease the memory and kernel reports it 2014-08-05 15:34:57 but when it is increased nothing happens. xl top lists it correctly, however kernel does not see it 2014-08-05 15:35:07 <^7heo_> katlogic: not that I know of. 2014-08-05 15:35:13 (need it for additional lxc instances running in dom0) 2014-08-05 15:35:26 where is size_t and FILE type defiend in musl ? 2014-08-05 15:35:49 <^7heo_> zxd: ask on #musl 2014-08-05 15:40:31 katlogic: it used to work 2014-08-05 15:40:46 darkfader: hm? 2014-08-05 15:41:06 i tried dding file into /dev/shm in hopes that buffer cache would ask xen for more memory 2014-08-05 15:41:11 instead it stupidly ooms 2014-08-05 15:41:20 kernel underreports memory, xl top lists the assigned number ... 2014-08-05 15:41:59 memory hotplug always gets broken not broken brocken not broken 2014-08-05 15:42:07 meh 2014-08-05 15:42:21 i'll try if it works on my host 2014-08-05 15:42:25 perhaps process needs to ask for the memory, it probably makes sense to not give leeway to buffer cache in this 2014-08-05 15:42:45 let my try with malloc/memset 2014-08-05 15:42:51 no if you do xl mem-set it has to show up right away 2014-08-05 15:43:49 perhaps it is related to having autobaloon=0 in xl.conf 2014-08-05 15:43:52 seems to work on mine 2014-08-05 15:43:53 no 2014-08-05 15:44:06 daveh0003:~# grep Total /proc/meminfo 2014-08-05 15:44:07 MemTotal: 32314896 kB 2014-08-05 15:44:08 daveh0003:~# xl mem-set Domain-0 34000 2014-08-05 15:44:12 daveh0003:~# grep Total /proc/meminfo 2014-08-05 15:44:12 MemTotal: 33338896 kB 2014-08-05 15:44:27 (XEN) Xen version 4.3.2 (buildozer@[unknown]) (gcc (Alpine 4.8.2) 4.8.2) debug=n Thu Jun 26 08:32:08 GMT 2014 2014-08-05 15:44:31 yours is? 2014-08-05 15:45:19 who did you query it in that format 2014-08-05 15:45:24 i have 4.3.1 anyway 2014-08-05 15:45:30 i'll try on a 4.3.2 host next 2014-08-05 15:45:49 xl dmesg 2014-08-05 15:48:01 waxh0012:~# grep MemTotal /proc/meminfo 2014-08-05 15:48:01 MemTotal: 1572856 kB 2014-08-05 15:48:01 waxh0012:~# xl mem-set Domain-0 3072 2014-08-05 15:48:01 waxh0012:~# grep MemTotal /proc/meminfo 2014-08-05 15:48:01 MemTotal: 2621432 kB 2014-08-05 15:48:18 (XEN) Xen version 4.3.2 (buildozer@[unknown]) (gcc (Alpine 4.8.2) 4.8.2) debug=n Thu Jun 26 08:32:08 GMT 2014 2014-08-05 15:48:58 i also have autoballoon=0 2014-08-05 15:50:07 do you set dom0_mem in the bootloader config? 2014-08-05 15:50:10 what kernel version? 2014-08-05 15:50:18 3.10.44-0-grsec over here 2014-08-05 15:51:03 nah, no dom0_mem in boot 2014-08-05 15:51:29 (XEN) Command line: iommu=1 2014-08-05 15:51:53 aha! after killing most dom0us, it suddenly increased 2014-08-05 15:52:05 probably related to fragmentation somewhere 2014-08-05 15:52:13 *most domUs 2014-08-05 15:52:32 i have "numa" in my options and "dom0_mem=2048", some of those could be related 2014-08-05 15:53:51 another weird thing, the kernel reports 400M less memory than assigned, is this normal thing? 2014-08-05 15:53:58 (the box is 16GB) 2014-08-05 15:54:02 meh, yes, i hate that too 2014-08-05 15:54:06 same for me 2014-08-05 15:54:14 that's since some time last year 2014-08-05 15:54:20 <^7heo_> katlogic: it's overcomitting probably. 2014-08-05 15:54:38 ^7heo_: didnt know xen has such a thing 2014-08-05 15:54:41 <^7heo_> linux has. 2014-08-05 15:54:46 <^7heo_> and probably, xen too. 2014-08-05 15:54:47 nah, i mean total memory 2014-08-05 15:54:50 hit badly when i only gave 300ish mb to dom0 2014-08-05 15:54:55 overcommit is unrelated :) 2014-08-05 15:55:14 <^7heo_> how else can you explain it? 2014-08-05 15:55:36 ^7heo_: it will underreport 400M-500M stable, no matter what i do 2014-08-05 15:55:49 it's more than the xen heap needs, too 2014-08-05 15:56:01 <^7heo_> ah damn 2014-08-05 15:56:02 (that is correctly accounted in xl, but that is more like 40mb) 2014-08-05 15:57:37 so I heard alpine is moving to systemd 2014-08-05 15:57:58 <^7heo_> haha. Funny. 2014-08-05 15:58:07 :) 2014-08-05 15:58:51 zxd, size_t type should be pulled in by stdlib.h or stddef.h 2014-08-05 16:00:28 <^7heo_> melville: he has the answer already I believe. 2014-08-05 16:04:10 ^7heo: yes but I didn't quite understand it , I think I understand is as in musl they hardcoded size_t to unsigned int with typset 2014-08-05 16:04:21 and in amd64 with unsigned long 2014-08-05 16:06:58 check bits/alltypes.h - this is internal header afaik 2014-08-05 16:07:21 typedef unsigned _Addr size_t 2014-08-05 16:07:56 and before that #define _Addr long on x86_64 2014-08-06 05:45:42 Hi, new to alpine. I saw in the 2.7.5 changelog that there is a package main/zfs-grsec, but apk search does not produce anything even after apk update && apk upgrade. Any pointer? 2014-08-06 08:59:05 zenny: it has been removed becaues of a kernel conflict 2014-08-06 08:59:25 but there are patches for 3.x 2014-08-06 12:40:04 kocka: do you have any links to the zfs patches for 3.x? 2014-08-06 12:52:23 capzulu: the ACF menu is stored in a session on the server, the only way to update it right now is to log off/on 2014-08-06 12:52:36 there is an exception is you use ACF to add a package 2014-08-06 12:52:57 since it knows packages have changed, it regeneates the menu 2014-08-06 12:53:10 (sorry the response took a few days, I've been afk) 2014-08-06 14:11:36 zenny: https://github.com/zfsonlinux/zfs/pull/2380 2014-08-06 15:45:21 kocka: Thanks for the link. BTW, do you have something like that for snapraid? 2014-08-06 19:45:04 zenny: nope :/ 2014-08-06 21:33:14 kocka: but the zfs patch in the link does not seem to work either with 3.0.2 release or zfs-0.6.3 :-( 2014-08-07 07:43:14 zelebar: i'm sorry to hear that :( 2014-08-07 11:21:52 hi 2014-08-07 11:22:32 i need help from a volunteer who can update documentation how to upgrade a v2.x install to v3.x 2014-08-07 11:40:44 "Include:Upgrading Alpine - HD v2.x to v3.x" - would be that proper wiki page? 2014-08-07 12:06:09 ncopa, I used sandbox for a draft: http://wiki.alpinelinux.org/wiki/Sandbox 2014-08-07 12:31:17 melville: looks good. Its --no-self-upgrade 2014-08-07 12:31:24 not --no-selfupgrade 2014-08-07 12:31:34 i can edit that 2014-08-07 12:31:57 ok 2014-08-07 12:32:45 copied that from irc logs, together with mistake ;) 2014-08-07 12:33:03 now, copy that to: http://wiki.alpinelinux.org/wiki/Upgrading_Alpine 2014-08-07 12:33:15 or merge it in 2014-08-07 12:33:21 thanks! 2014-08-07 12:41:55 http://wiki.alpinelinux.org/wiki/Upgrading_Alpine 2014-08-07 12:41:56 done 2014-08-07 13:10:25 ncopa: busy with new www? 2014-08-07 14:04:13 melville: thanks! 2014-08-07 14:04:23 clandmet1r: not really. busy with other stuff 2014-08-07 14:39:57 Is there some known Issue with LVM on 3.0.2? I get OOM Error Messages when creating/deleting lv's, although everything seems to work... http://pastebin.com/rDudAiht 2014-08-07 14:42:10 yeah 2014-08-07 14:43:07 http://bugs.alpinelinux.org/issues/3107 2014-08-07 14:44:06 the worst thing that can do is make the lvm code in memory swap out 2014-08-07 14:44:33 that only happens if you: 1) are low on memory, 2) have swap enabled 2014-08-07 14:44:59 i suppose we should fix it though 2014-08-07 14:45:35 ncopa: btw i got those errors on debian/arm (32bit) for a long time, it seems to be a general lvm stoopid 2014-08-07 14:45:52 of course there i'm likely to be low on memory and have swap 2014-08-07 14:46:11 i think that is what they are trying to do 2014-08-07 14:46:19 make sure that lvm is never swapped out 2014-08-07 14:46:28 in case it will not be able to swapin again 2014-08-07 14:47:08 darkfader: could you please comment on #3107 and possibly have a link that refers to the same issue on debian/arm? 2014-08-07 14:50:31 ok i'll check if there is something reported for debian 2014-08-07 14:51:11 and that would be a good explanation, yes 2014-08-07 14:51:33 pitfalls of putting critical components in userspace because userspace is inherently awesome :) 2014-08-07 14:56:09 Thanks for the Info Guys! I will disable swap for the time being 2014-08-07 15:00:36 tze: it is only a problem when starting/stopping the lvm i think 2014-08-07 15:07:31 ncopa: i got it on every lvcreate i think 2014-08-07 15:07:41 can't tell for sure but i'll include info in the bug 2014-08-07 15:07:57 tze: it's informal imo 2014-08-07 15:10:32 right 2014-08-07 15:10:35 i mean lvcreate 2014-08-07 15:10:43 you dont do that too often i suppose 2014-08-07 15:11:56 err 2014-08-07 15:12:34 depends 2014-08-07 15:12:42 on some systems i do that daily 2014-08-07 15:13:02 i.e. vm hosts 2014-08-07 15:13:14 of course i learned to ignore the message 2014-08-08 07:21:13 Does alpine use grsec? 2014-08-08 07:21:32 Ah yes 2014-08-08 07:21:41 I was looking at the wrong part of the Wiki 2014-08-08 07:24:30 Does alpine use position independent executables? 2014-08-08 07:26:13 fowlslegs, yes, and yes 2014-08-08 07:26:32 we build *everything* with PIE, relro, and bind now 2014-08-08 07:26:37 and stack-protector 2014-08-08 07:27:44 Do you use -fstack-protector-all for SSP? 2014-08-08 07:28:30 yes 2014-08-08 07:28:46 considering to use -fstack-protector-strong after migration to gcc-4.9.x 2014-08-08 07:33:05 sadly stack-protector-strong does not do the right thing 2014-08-08 07:33:20 Can you expand dalias 2014-08-08 07:33:28 yeah 2014-08-08 07:33:58 "Like -fstack-protector but includes additional functions to be protected — those that have local array definitions, or have references to local frame addresses" 2014-08-08 07:34:01 I read Google developed it, but still uses all for their personal uses 2014-08-08 07:34:37 having a local array definition or referencing local frame addresses is not all that risks overwriting the return address 2014-08-08 07:35:02 dalias, I was asking to explain what functions, etc., "strong" doesn't protect that "all" does 2014-08-08 07:35:31 can also wonder why they dont do -fstack-protector=strong|all instead of introducing another -fstack-protector-another-try 2014-08-08 07:35:34 void foo(int *p) { p[-4] = evil; } 2014-08-08 07:35:58 foo(&local); 2014-08-08 07:36:08 has a good chance of clobbering foo's return address 2014-08-08 07:36:41 the condition for needing a canary check is "does the function make any writes via a pointer that can't be proven valid?" 2014-08-08 07:37:17 (and "any writes via a pointer" would be a sufficient check) 2014-08-08 07:37:47 right now, stack-protector-all is really wasteful and can cripple performance 2014-08-08 07:37:59 by adding checks to pure arithmetic functions that should be very fast 2014-08-08 07:39:08 fowlslegs: -fstack-protector-all adds checks to functions that does not need it 2014-08-08 07:39:25 thats what -strong was supposed to solve 2014-08-08 07:40:20 "pure arithmetic functions" like: int add(int a, int b) { return a+b; }; 2014-08-08 07:41:20 Okay. Have Alpine devs considered writing a patch that removes these checks from the "all" option? 2014-08-08 07:42:00 there is such option in gcc-4.9, which is called -fstack-protector-strong 2014-08-08 07:42:10 google already did that job :) 2014-08-08 07:45:20 Ah, but doesn't strong not check certain fns as dalia just pointed out? 2014-08-08 07:45:48 right 2014-08-08 07:45:54 strong misses functions that do need checks 2014-08-08 07:46:05 we should try to get this fixed in gcc 2014-08-08 07:46:20 a good starting point would be "any write through pointer" 2014-08-08 07:46:28 then they could optimize it further 2014-08-08 07:46:41 yeah... too bad there is so much other broken software we also need to fix... :) 2014-08-08 07:46:47 "stronger" 2014-08-08 07:46:50 e.g. this function does not need a canary: 2014-08-08 07:47:54 int foo() { int a[3] = {1,2,3}; int i; for (i=0; i<1000; i++) a[rand()%3] ^= rand() ^ a[rand()%3]; return a[0]; } 2014-08-08 07:48:19 it writes via pointers, even in a fairly complex pattern 2014-08-08 07:48:29 but the writes are trivially provable to be within the bounds of the array 2014-08-08 07:52:49 Have you looked at the patches Hardened Gentoo uses for GCC? 2014-08-08 07:55:30 no 2014-08-08 07:55:38 i have 2014-08-08 07:55:47 and basically, that is what we use 2014-08-08 08:05:30 Are there any major security differences, as far as you know, between Gentoo hardened uClibc and Alpine? Do you use RBAC as well? 2014-08-08 08:15:58 i'm not aware of anything major 2014-08-08 08:16:35 there are some benefits of running your own built grsecured kernel over using a prebuilt distro-shipped 2014-08-08 08:16:54 i dont use RBAC personally 2014-08-08 08:18:27 well, code quality is much better in musl libc than uclibc 2014-08-08 08:18:43 so that might make a difference 2014-08-08 08:22:59 What cryptographic algorithms do you use? 2014-08-08 08:24:02 Sorry to ask so many questions. I just don't know how I would find out otherwise. As you stated on your wiki there is much missing. 2014-08-08 08:25:11 E.g., for password hashing. 2014-08-08 08:26:44 oh thats another thing 2014-08-08 08:27:28 iirc think we have support for sha512/sha256/bcrypt (aka blowfish) 2014-08-08 08:27:38 and i think we patched uclibc for those too 2014-08-08 08:27:45 dunno if gentoo uclibc does that 2014-08-08 08:28:22 i can also mention that the way musl handled the one sec issue it had 2014-08-08 08:28:34 i have never seen anything like it in my years with uclibc 2014-08-08 08:35:29 Is there support for two/threefish? 2014-08-08 08:38:00 "At this point, though, I'm amazed it's still being used. If people ask, I recommend Twofish instead." - Schneier on blowfish 2014-08-08 08:39:22 I dislike the Linux community's larger trend of staying behind the curve when it comes to using broken protocols like MD5. 2014-08-08 08:44:41 not that i know 2014-08-08 08:48:14 Is this something that has to be built into/ patched onto musl? Can it be ported from glibc? On one hand, when cryptographic algorithms have lasted decades without being broken, that can say something about its security, on the other hand the longer they are around for the greater chance there is that they have been broken privately. 2014-08-08 08:49:10 It seems like the best choices are the ones that have been around for an intermediate amount of time without being broken and are popular--implying they have been studied extensively. 2014-08-08 08:49:35 Another thing about modern algorithms is they have increasingly taken side-channel attacks into account. 2014-08-08 08:50:37 E.g., Threefish uses no S-boxes or other table lookups in order to avoid cache timing attacks. 2014-08-08 08:51:03 Blowfish is susceptible to these types of attacks. 2014-08-08 08:54:06 However, there have been some strong cryptanalytic attacks published on Threefish. 2014-08-08 08:56:47 So I found this list of crypto libraries that may be usable by musl http://wiki.musl-libc.org/wiki/Alternative_libraries#Crypto 2014-08-08 08:56:53 usable with* 2014-08-08 09:12:18 Do you use the Linux IMA? 2014-08-08 09:44:15 no, have not used linux IMA 2014-08-08 12:10:13  2014-08-08 12:38:12 <^7heo> Moin here. 2014-08-08 12:38:20 <^7heo> I have problems with alsa 2014-08-08 12:38:35 <^7heo> anyone also had problems with alsa out of the box? 2014-08-08 12:38:57 depends what kind of problems 2014-08-08 12:39:07 i got it running here with mpd 2014-08-08 12:39:49 <^7heo> problems of buffer overrun and very poor sound quality 2014-08-08 12:50:30 <^7heo> okay 2014-08-08 12:50:33 <^7heo> alsa isn't the problem 2014-08-08 12:50:34 <^7heo> vlc is. 2014-08-08 12:50:47 <^7heo> sorry for the noie. 2014-08-08 12:50:50 <^7heo> noise even 2014-08-08 12:50:55 np 2014-08-08 12:51:02 we are used to it :) 2014-08-08 12:51:20 <^7heo> You mean, from me, or the community in general? 2014-08-08 12:53:09 <^7heo> mpg123 is actually very fine. 2014-08-08 12:54:28 in general 2014-08-08 12:54:56 thats what this channel is for. 2014-08-08 12:55:01 <^7heo> yeah right 2014-08-08 12:55:30 <^7heo> btw, I have problems with chrony, any idea how to perform the command "ntpdate pool.ntp.org" with chrony? 2014-08-08 13:04:05 <^7heo> damn, I keep getting 501 errors. 2014-08-08 13:06:39 <^7heo> btw, where is the alpine linux github? 2014-08-08 13:07:00 <^7heo> https://github.com/alpinelinux ? 2014-08-08 13:07:08 https://github.com/alpinelinux 2014-08-08 13:54:32 <^7heo> I believe that the default chrony config is faulty. 2014-08-08 13:56:31 <^7heo> about the github 2014-08-08 13:56:43 <^7heo> should I fork/request? 2014-08-08 13:56:57 <^7heo> or do you give access to the community members? 2014-08-08 14:06:27 fork and pr 2014-08-08 14:52:32 <^7heo> clandmeter: okay 2014-08-08 15:43:26 <^7heo> I'm missing a command in apk 2014-08-08 15:43:58 <^7heo> apk list 2014-08-08 15:44:01 <^7heo> also 2014-08-08 15:44:10 <^7heo> two other commands 2014-08-08 15:44:37 <^7heo> (apk info == apk list?) 2014-08-08 15:46:22 <^7heo> I would like apk find (find file in package) 2014-08-08 15:47:27 <^7heo> and apk ls (list files in package) 2014-08-08 16:51:20 hi 2014-08-08 16:54:16 I have this program http://ideone.com/QCNgoq , running with gdb but for some reason when stepping into to the foo function I don't see the args passed to it like they say in this C book example 2014-08-08 16:54:40 c book example shows #0 foo (x={month = 10, day = 11, year = 2004}) at p18-5.c:13 I see foo (x=...) at listmain.c:9 2014-08-08 16:54:47 I don't know if it's related to PAX 2014-08-08 16:54:55 kernel protection stuff grsec? 2014-08-08 17:09:38 never mind I see it now 2014-08-08 17:09:41 asked in gdb 2014-08-08 17:09:47 set print frame-arguments all 2014-08-08 17:12:53 have a different problem now when restarting a program gdb hangs 2014-08-08 17:13:20 err 2014-08-08 17:13:24 it works just takes time I guess 2014-08-09 01:35:13 <^7heo__> btw, does anyone has had problems with sound on firefox? 2014-08-09 09:45:35 o hai 2014-08-09 09:45:49 Tell me, am I being stupid, or does apk not have a way of ignoring deps? 2014-08-09 09:46:07 I'm experimenting with Alpine and want to remove openrc from it 2014-08-09 09:46:57 I can, of course, rename my own /sbin/rc to something else, but I as a control freak, I want my system to only contain the one service manager I actually use :) 2014-08-09 09:54:01 Also there's no provides= in the apkbuilds, I see... bummer 2014-08-09 09:59:56 create stub package for openrc? would this work? 2014-08-09 10:05:41 It would, but it would have to have the same name as openrc, right? 2014-08-09 10:05:55 As there's no provides= 2014-08-09 10:06:57 i mean empty openrc package with higher number or meta package that depends on your rc 2014-08-09 10:08:48 yeah, I get it. I guess it will have to do. 2014-08-09 10:17:57 openfbtd, there is also this http://wiki.alpinelinux.org/wiki/APKBUILD_Reference#replaces 2014-08-09 10:18:17 not sure how useful it is 2014-08-09 10:20:01 Hmmmm 2014-08-09 10:20:03 Hmmmmmmmm 2014-08-09 10:20:07 That may work. 2014-08-09 10:20:13 Depending on how apk works 2014-08-09 10:20:15 Thx 2014-08-09 10:20:18 I'll try it. 2014-08-09 11:03:56 Oh well, at least this pushed me to write a makefile for spark-rc. 2014-08-09 11:38:40 Arrr, I'm so used to bindirs being merged! I need to fix my initscripts because they all point to /usr/bin >_> 2014-08-09 11:39:26 <^7heo> openfbtd: why not use unionsfs? 2014-08-09 11:40:26 That's a solution, I guess 2014-08-09 11:42:08 <^7heo> I mean, I'm a huge fan of unionfs for whatever it can be. 2014-08-09 11:42:50 Also the tar port makes a /bin/tar -> /usr/bin/tar link 2014-08-09 11:43:09 And if you try merging the bindirs, it backfires :D 2014-08-09 11:43:17 <^7heo> yeah 2014-08-09 11:43:28 <^7heo> however, with unionfs, you just mask one with the other. 2014-08-09 11:43:37 <^7heo> and if you do it right, you mask the link 2014-08-09 11:43:47 <^7heo> ;0 2014-08-09 11:43:57 I don't get why the link is even there 2014-08-09 11:44:43 <^7heo> so it is "more compatible" when people call /bin/tar? 2014-08-09 11:44:52 In the meantime, I have something like https://fleshless.org/pages/spark.html, but Alpine 2014-08-09 11:45:09 But that's retarded. People should not call /bin/tar ever in the history of ever 2014-08-09 11:45:18 That's what's PATH is for 2014-08-09 11:45:28 That's what PATH is for* 2014-08-09 11:45:52 <^7heo> problem is, arch is broken by design 2014-08-09 11:46:02 <^7heo> so doing it with arch will be sucky, no matter what. 2014-08-09 11:46:03 And if you hardcode any paths except for maybe /usr/bin/env, you should die in a fire. 2014-08-09 11:46:37 Any executable paths, obviously. 2014-08-09 11:46:44 <^7heo> actually, even /usr/bin/env 2014-08-09 11:46:53 <^7heo> the ONLY safe path to point at is /bin/sh 2014-08-09 11:46:58 <^7heo> full stop. 2014-08-09 11:46:59 You should just call env, yeah. 2014-08-09 11:47:06 <^7heo> exactky. 2014-08-09 11:47:09 <^7heo> exactly even 2014-08-09 11:47:32 I have /usr/bin/env in my scripts, but that should change when I'm not too lazy to push new releases of everything ever. 2014-08-09 11:47:47 <^7heo> I always start ANY script with "/bin/sh" 2014-08-09 11:47:57 <^7heo> and I check what's available and call accordingly. 2014-08-09 11:48:01 My scripts explicitly depend on bash a lot of the time 2014-08-09 11:48:11 <^7heo> I hate the bashisms 2014-08-09 11:48:17 <^7heo> most are implementable in sh 2014-08-09 11:48:21 Well that's a preference. 2014-08-09 11:48:39 <^7heo> and if it's not, you have external tools, or you can distribute your own C code for it. 2014-08-09 11:48:52 I just went “ok my scripts depend on bash, so either install it or not use my scripts” 2014-08-09 11:48:57 at some point. 2014-08-09 11:49:22 <^7heo> personally, I'm fine with people using bash, as long as it is REPLACABLE in distributed products. 2014-08-09 11:49:53 There is no point of using BASH specifically if you write POSIX shell scripts 2014-08-09 11:50:13 I write bash scripts because I actually _use_ bash specific features. 2014-08-09 11:50:38 <^7heo> yeah 2014-08-09 11:50:40 <^7heo> problem is 2014-08-09 11:50:46 <^7heo> you're part of a TINY minority 2014-08-09 11:50:55 <^7heo> arch people write bash scripts calling /bin/sh 2014-08-09 11:51:07 <^7heo> gog people write bash scripts where they don't do any bashisms 2014-08-09 11:51:08 <^7heo> etc. 2014-08-09 11:51:14 I hate when people do that. 2014-08-09 11:51:16 <^7heo> and they all happily "support Linux" 2014-08-09 11:51:22 Call /bin/sh for bash, I mean. 2014-08-09 11:51:35 If you call /bin/sh, DON'T USE BASHISMS 2014-08-09 11:51:36 <^7heo> that's my reason for ditching arch, back in the days 2014-08-09 11:51:57 $ head -1 /bin/makepkg 2014-08-09 11:51:57 #!/usr/bin/bash 2014-08-09 11:51:59 btw 2014-08-09 11:52:00 <^7heo> they did that on packages, so, once I replaced /bin/sh, and upgraded my distro, everything broke 2014-08-09 11:52:05 <^7heo> I reported the bug 2014-08-09 11:52:06 That was cleaned up at some point 2014-08-09 11:52:09 <^7heo> wontfix 2014-08-09 11:52:14 <^7heo> I complained 2014-08-09 11:52:23 <^7heo> I got insulted, then banned. 2014-08-09 11:52:37 <^7heo> kind of a straightforward way to add insult to injury 2014-08-09 11:52:58 Arch community can suck a lot. Often. 2014-08-09 11:53:03 <^7heo> Always. 2014-08-09 11:53:06 <^7heo> Not just often. 2014-08-09 11:53:17 <^7heo> It's a puddle of arrogant noob pricks. 2014-08-09 11:53:23 <^7heo> hardly any skill there. 2014-08-09 11:53:31 <^7heo> They just masturbate at their own bugs. 2014-08-09 11:53:47 You are either like me and have really thick skin (raised by the Internet in general and imageboards in particular) or don't deal with arch people. 2014-08-09 11:54:14 <^7heo> if I could nuke ANY distribution on earth, I wouldn't nuke ubuntu or anything that most geeks consider wrong like red hat. 2014-08-09 11:54:15 But then I'm also not a nice person. At all. 2014-08-09 11:54:19 <^7heo> I would nuke Arch. 2014-08-09 11:54:33 <^7heo> And if I could make the more Arch developers suffer, I'd even accept to suffer for it. 2014-08-09 11:55:05 <^7heo> I don't know about my skin 2014-08-09 11:55:12 <^7heo> but there is one thing I hate above all 2014-08-09 11:55:23 <^7heo> when arrogant people cause major harm because of their stupidity. 2014-08-09 11:55:46 Thing is, so far Arch is the only distro with a huge community (-> huge relatively tested repos) that allows for a lot of tinkering. 2014-08-09 11:55:49 <^7heo> and when they do that rightfully in the name of "majority" 2014-08-09 11:55:55 <^7heo> of the "Majority" sorry 2014-08-09 11:55:59 Technically Arch is mostly fine these days. 2014-08-09 11:56:06 But fuck that community. 2014-08-09 11:56:08 <^7heo> systemd isn't fine 2014-08-09 11:56:14 systemd is easily removed 2014-08-09 11:56:23 <^7heo> and the developers are leading the "fuck" 2014-08-09 11:56:35 <^7heo> I won't contribute to anything close to arch 2014-08-09 11:56:51 I've not encountered a single thing in Arch that is not removed relatively easily if you know what you're doing. 2014-08-09 11:56:53 <^7heo> these minds are just ruining open source. 2014-08-09 11:57:01 <^7heo> well 2014-08-09 11:57:23 <^7heo> replacing /bin/sh by !bash; when the pacman scripts depend on it, is impossible. 2014-08-09 11:57:27 <^7heo> Even if you know what you're doing. 2014-08-09 11:57:47 That was cleaned up. 2014-08-09 11:57:55 Scripts that use bash call bash these days 2014-08-09 11:58:07 Explicitly 2014-08-09 11:58:09 <^7heo> You told it already. I didn't knew before, because I stay away from arschlinux. 2014-08-09 11:58:26 <^7heo> but it's just to show the sheer amount of shit they can expose you to. 2014-08-09 11:58:29 <^7heo> I don't want that. 2014-08-09 11:58:55 I'm looking at Alpine to replace the base for Spark actually. 2014-08-09 11:59:03 Because so far I really like what I see. 2014-08-09 11:59:14 <^7heo> Alpine? 2014-08-09 11:59:15 But that's not going to happen soon 2014-08-09 11:59:17 <^7heo> Imho it's sane. 2014-08-09 11:59:22 It is. 2014-08-09 11:59:36 <^7heo> It's not active enough, not mature enough, but it's getting there. 2014-08-09 11:59:40 How is it for a full desktop? 2014-08-09 11:59:44 <^7heo> Great 2014-08-09 11:59:57 <^7heo> I can't run steam because it's linked againt the gbloatc, but that's about it. 2014-08-09 11:59:59 I want to both work and play on my machines 2014-08-09 12:00:06 <^7heo> oh and vlc sucks a lot in Alpine, but mplayer is totally fine. 2014-08-09 12:00:06 Yep that's a full stop 2014-08-09 12:00:18 I'm a heavy gamer >_> 2014-08-09 12:00:22 <^7heo> Same 2014-08-09 12:00:29 <^7heo> so heavy that I actually use windows for that. 2014-08-09 12:00:39 <^7heo> only approx 20% of my games work on linux. 2014-08-09 12:00:43 Ah. I'm not ready to torture myself with Windows. 2014-08-09 12:00:54 So I'm torturing mysellf with WINE 2014-08-09 12:00:55 :D 2014-08-09 12:01:00 <^7heo> it's not a torture if you don't do ANYTHING with it but booting it like you would start your xbox. 2014-08-09 12:01:11 Yeah. The problem if you need to reboot 2014-08-09 12:01:17 And lose all your workflow 2014-08-09 12:01:17 <^7heo> Nah I don't. 2014-08-09 12:01:22 <^7heo> That's the trick. 2014-08-09 12:01:24 <^7heo> Xen ;) 2014-08-09 12:01:28 Ah. A separate gaming machine? 2014-08-09 12:01:30 Aaaah 2014-08-09 12:01:37 GPU passthrough? 2014-08-09 12:01:41 <^7heo> exact ;) 2014-08-09 12:01:52 I've been looking at that at some point 2014-08-09 12:01:56 <^7heo> one core on my linux, 'cause I don't need much, with a small GPU 2014-08-09 12:02:01 <^7heo> and the rest goes to windows. 2014-08-09 12:02:15 But these days most of the games I want to play work on linux either through WINE or natively 2014-08-09 12:02:22 <^7heo> yeah 2014-08-09 12:02:24 So I'll probably never get to it 2014-08-09 12:02:36 <^7heo> however, they require bash, glibc, and various shitty libraries. 2014-08-09 12:02:47 <^7heo> so I really prefer to keep the shit in windows, if you get my meaning. 2014-08-09 12:02:59 So far it looks like I'm stuck on Arch forever >_> 2014-08-09 12:03:09 <^7heo> nah you're not 2014-08-09 12:03:18 Alpine looks really nice for my smaller machines 2014-08-09 12:03:25 <^7heo> getting IOMMU with vt-d is not easy, but doable. 2014-08-09 12:03:34 I'll probably run it on my netbook. 2014-08-09 12:03:53 Used to run OpenBSD on it actually 2014-08-09 12:03:59 <^7heo> I use alpine on a i7-4810MQ with 16GB of ram and a GTX 860M :P 2014-08-09 12:04:00 Alpine seems like a better solution 2014-08-09 12:04:07 <^7heo> so it's not like it's ONLY for small machines. 2014-08-09 12:04:27 It's not. It's just that in my case it'd be nice for smaller machines 2014-08-09 12:04:34 <^7heo> I like Alpine and NetBSD 2014-08-09 12:04:36 <^7heo> not much more. 2014-08-09 12:04:38 Because I don't want windows for games :) 2014-08-09 12:04:41 Even in a vm 2014-08-09 12:05:04 I mean, it's an ok solution. But not ok enough for me. 2014-08-09 12:05:13 <^7heo> I don't care about that, because my VM is subject to the IPtables rules of my host 2014-08-09 12:05:19 <^7heo> of all the "jailing" 2014-08-09 12:05:30 If a game doesn't work on linux in some way or another... fuck it, I'm not playing it 2014-08-09 12:05:30 <^7heo> and nothing ever happens there excepted the installation of drivers and games. 2014-08-09 12:05:40 <^7heo> borderlands one is great tho. 2014-08-09 12:05:53 Both bthe BL games work on WINE :) 2014-08-09 12:05:55 the* 2014-08-09 12:06:04 <^7heo> ah, I thought only the second was. 2014-08-09 12:06:13 Nope, I played through both on WINE. 2014-08-09 12:06:43 <^7heo> okay 2014-08-09 12:06:50 <^7heo> point is 2014-08-09 12:07:02 <^7heo> I'm a sysadmin. I spend my days making broken software work. 2014-08-09 12:07:05 ACTION runs alpine as host OS, windows xp as a xen guest + 3d GFX PCI passthrough 2014-08-09 12:07:11 if you're lucky the passthrough works really well 2014-08-09 12:07:16 <^7heo> yeah 2014-08-09 12:07:26 <^7heo> So when I go home, the LAST thing I do is to solve problems with broken software. 2014-08-09 12:07:55 <^7heo> Therefore, a stock windows with steam + GPU drivers is exactly what I need. 2014-08-09 12:08:10 <^7heo> and if I ever need to compile stuff 2014-08-09 12:08:22 <^7heo> I can either shutdown the windows vm for a while 2014-08-09 12:08:31 <^7heo> or compile remotely on some servers 2014-08-09 12:08:34 <^7heo> or on my laptop 2014-08-09 12:08:37 <^7heo> and that's good enough 2014-08-09 12:08:42 <^7heo> I don't compile THAT often. 2014-08-09 12:08:52 That's the curious thing. I see a lot of people take that position 2014-08-09 12:08:54 14:04 < ^7heo> I like Alpine and NetBSD 2014-08-09 12:08:54 14:04 < ^7heo> not much more. 2014-08-09 12:08:57 the same here 2014-08-09 12:09:02 <^7heo> :) 2014-08-09 12:09:11 But I'm also a sysadmin. And I tinker with software ALL the time 2014-08-09 12:09:15 At home and at work 2014-08-09 12:09:19 I just love it :) 2014-08-09 12:09:28 <^7heo> I really, REALLY want to stop tinkering when I play. 2014-08-09 12:09:52 <^7heo> nothing's more frustrating than when you want to have it working now with a friend, you got the friend, the pizzas, the beers, and... 2014-08-09 12:10:00 <^7heo> BAM, no sound, or missing dep, or whatever. 2014-08-09 12:10:00 It's often “ok I wanna play this ga-REBIUILDING WINE WITH SEVEN HUNDRED PATCHES NOW” 2014-08-09 12:10:03 <^7heo> and then? 2014-08-09 12:10:19 I gave no friends. 2014-08-09 12:10:19 <^7heo> http://xkcd.com/349/ 2014-08-09 12:10:24 <^7heo> then that ^ 2014-08-09 12:10:25 ^7heo: for friends & beer i think xbox works better :/ 2014-08-09 12:10:30 openfbtd is archetypal pc gamer 2014-08-09 12:10:31 <^7heo> katlogic: sorry, I can't. 2014-08-09 12:10:36 <^7heo> katlogic: I need steam for my games. 2014-08-09 12:10:36 Also I would actually set up the game before the friend comes there. 2014-08-09 12:11:05 ACTION never bought a game, so has a bit skewed perspective 2014-08-09 12:11:08 <^7heo> yeah, you're the careful type. 2014-08-09 12:11:11 It would be extremely stupid of me not to make sure the game works BEFORE inviting someone to play it with me 2014-08-09 12:11:17 ^7heo: there are consoles :) 2014-08-09 12:11:21 <^7heo> I'm more of the spontaneous type. 2014-08-09 12:11:24 and homebrew! 2014-08-09 12:11:31 I'm using xbox when i want to play 2014-08-09 12:11:40 <^7heo> again, I don't want to. 2014-08-09 12:11:43 <^7heo> PC gaming FTW. 2014-08-09 12:11:44 People tell me I'm a german who was born somewhere else >_> 2014-08-09 12:11:48 <^7heo> Consoles are bound to die. 2014-08-09 12:12:06 just put in dvd & play, xbox is awesome how dumb-proof it is for computer illiterates 2014-08-09 12:12:27 <^7heo> yeah but I want to be able to change the components 2014-08-09 12:12:30 or drunk people; same thing 2014-08-09 12:12:31 ^7heo, Valve are doing good work on trying to kill consoles 2014-08-09 12:12:32 <^7heo> tweak the graphics 2014-08-09 12:12:33 <^7heo> etc. 2014-08-09 12:12:37 <^7heo> openfbtd: yes 2014-08-09 12:12:45 <^7heo> openfbtd: and I'm throwing a lot of money at them for that 2014-08-09 12:12:50 By replacing them with small nice pcds 2014-08-09 12:12:52 pcs* 2014-08-09 12:12:55 <^7heo> yes 2014-08-09 12:13:01 <^7heo> next stop after the living room? 2014-08-09 12:13:05 ^7heo: well, sure, thats what pc is for. it's just that using pc as a console (ie having separate gaming pc) sounds like waste of money. 2014-08-09 12:13:06 <^7heo> your smartblets. 2014-08-09 12:13:07 Yeah, I've bought everything I ever wanted on Steam so far 2014-08-09 12:13:13 So can't really throw a lot more at them 2014-08-09 12:13:24 Sales are no fun anymore 2014-08-09 12:13:26 <^7heo> katlogic: it's not a waste of money if you have the money ;) 2014-08-09 12:13:37 <^7heo> openfbtd: how many games do you have/ 2014-08-09 12:13:39 <^7heo> ? 2014-08-09 12:13:41 “Oooh, a sale, lots of games to b-wait a second, I have all these” 2014-08-09 12:13:54 haha, must sucks to be rich/buy games 2014-08-09 12:13:58 Closing to 300 now. I'm picky. 2014-08-09 12:14:04 <^7heo> I'm at 666 atm. 2014-08-09 12:14:13 <^7heo> And I add the new ones in my inventory to not change the count. 2014-08-09 12:14:22 One of my coworker has closer to 800 2014-08-09 12:14:32 <^7heo> I think I will be close to 800 next steam sales 2014-08-09 12:14:35 ^7heo: every time you buy some new game, you have to gift one away? :) 2014-08-09 12:14:40 Also playing a lot of Hearthstone lately. Threw some money at that >_> 2014-08-09 12:14:42 <^7heo> but right now 666 is a good count 2014-08-09 12:14:46 ACTION seen one guy with 1337 count do exactly that 2014-08-09 12:14:51 <^7heo> katlogic: not everytime. 2014-08-09 12:15:06 katlogic, you can't give away a game you've added to the collection 2014-08-09 12:15:21 And you can't play games from your inv 2014-08-09 12:15:26 <^7heo> unless you buy 4packs :P 2014-08-09 12:15:46 One of them you get in your library 2014-08-09 12:15:47 <^7heo> but personally I prefer 6 packs. 2014-08-09 12:15:50 <^7heo> ACTION hides 2014-08-09 12:15:59 <^7heo> openfbtd: yeah ofc 2014-08-09 12:16:06 <^7heo> okay, I gtg anyway 2014-08-09 12:16:08 <^7heo> I'm starving 2014-08-09 12:16:12 <^7heo> my dog needs to walk 2014-08-09 12:16:15 <^7heo> and I have stuff to do 2014-08-09 12:16:17 <^7heo> so... 2014-08-09 12:16:23 <^7heo> t'was nice, but bye for now ;) 2014-08-09 12:16:24 <^7heo> o/ 2014-08-09 12:17:12 cya 2014-08-09 12:19:34 openfbtd: yeah, only games you get twice (ie single buy and package buy) can be gifted 2014-08-09 12:19:38 for the rest, contact steam support 2014-08-09 12:19:46 yay for feudal overlords of app markets 2014-08-09 14:25:21 looks nice http://www.legitreviews.com/mad-catz-ships-s-t-r-k-e-wireless-keyboard_148290 2014-08-09 14:25:31 one keyboard for multiple devices 2014-08-11 14:33:33 <^7heo> yo people 2014-08-13 10:41:06 hoi tg 2014-08-13 10:41:08 sup there? 2014-08-14 15:05:38 anyone know the package name that contains ypbind? 2014-08-14 23:48:15 Hi, can anyone tell me how to install kvm and openvswitch to alipine? 2014-08-14 23:51:40 Any pointers to install kvm and openvswith to alipine be appreciated! 2014-08-15 08:52:52 whats the minimum size of the alpine linux usb stick? 2014-08-15 08:59:20 depends 2014-08-15 08:59:41 if no custom iso, check size of alpine-mine 2014-08-15 08:59:53 alpine-mini 2014-08-15 09:03:29 Is it possible to install kvm and openswitch to alpine? 2014-08-15 09:03:45 kvm yeah 2014-08-15 09:04:09 our infa runs partially on kvm 2014-08-15 09:04:42 ovs i dont know the status. 2014-08-15 09:05:43 @shafire @clandmeter: is there any documents related to the kvm setup in alpine? ovs is in v 2.3.0 and works well in debian. 2014-08-15 09:08:30 zenny: ovs is only available in uclibc edge testing 2014-08-15 09:09:49 err also in musl edge/testing 2014-08-15 09:11:26 zenny: if its not in our wiki, i guess you need to take it from another source. 2014-08-15 09:11:59 clandmeter: thanks, but what makes me surprised is the edge/testing is at v 3.0.0 whereas stable is at v 3.0.2, a bit confused about versioning of alpine. 2014-08-15 09:12:52 clandmeter: any links handy for ovs+musl in edge and kvm? 2014-08-15 09:13:02 where do you get the version number from? 2014-08-15 09:14:39 zenny: i would install alpine stable 3.0.x and pin testing repo so you can try ovs out. 2014-08-15 09:15:48 from the download links. http://nl.alpinelinux.org/alpine/edge/releases/x86_64 http://nl.alpinelinux.org/alpine/v3.0/releases/x86_64 2014-08-15 09:16:00 zenny: kvm uses qemu. and we use libvirt to manage it. 2014-08-15 09:16:23 zenny: did you already isntall alpine? 2014-08-15 09:18:37 clandmeter: I tried install alpine-xen to a kvm instance, but keeps on giving ""BUG: #CPU1 idle for n seconds". Installed 3.0.2 2014-08-15 09:19:23 Tried to install edge and it crashes after network setup with setup-alpine. New to alpine ;-) 2014-08-15 09:21:31 Also tried to build zfs (https://github.com/zfsonlinux/zfs/pull/2380), but without success. 2014-08-15 09:22:59 btrfs :-) 2014-08-15 09:24:28 shafire: btrfs does not seem comparable to zfs in features and performance as far as I read and experimented with. 2014-08-15 09:30:09 for me it is the opposite 2014-08-15 09:31:57 shafire: would you mind sharing your experience with btrfs? 2014-08-15 09:32:09 zfs was slow after having a lot of subvolumes 2014-08-15 09:33:24 is there some versioning/timestamping of apkindex files? 2014-08-15 09:33:33 and for me: btrfs is stable and in kernel, zfs is not in kernel 2014-08-15 09:37:01 shafire: kernel inclusion is a known issue, but what scares everyone is the institution behind the btrfs development. 2014-08-15 09:37:37 "everyone" 2014-08-15 09:37:51 same with kernel 2014-08-15 09:39:25 zenny: try http://nl.alpinelinux.org/alpine/v3.0/releases/x86_64/alpine-3.0.3-x86_64.iso 2014-08-15 09:39:38 and you should be able to get kvm working 2014-08-15 09:40:02 after that you can pin edge repo and isntall ovs 2014-08-15 09:40:44 pinning repo: @testing http://nl.alpinelinux.org/alpine/edge/testing/ 2014-08-15 09:40:58 apk update && apk add openvswitch@testing 2014-08-15 09:41:25 clandmeter: thanks i am upgrading alpine 3.0.2 to 3.0.3 and shall pin the repos as you suggested. Quite exciting. 2014-08-15 09:42:04 zenny: be carefull with pinned repo's. sometimes edge pkgs are not compat with stable. 2014-08-15 09:42:50 abi changes in edge will break pkgs in stable when pinned. 2014-08-15 09:44:28 shafire: why kernel? 2014-08-15 09:47:05 clandmeter: I cloned the installed alpine to test with edge. Let's see how it goes. 2014-08-15 09:47:44 hrm, the web shows different timestamp on the apkfile. timezone perhaps? I wonder if I should use etag instead, building fs images and I want to know what exactly it was built from 2014-08-15 09:48:09 CcxCZ: the timestamp is in apkindex 2014-08-15 09:49:02 atleast i think it is 2014-08-15 09:49:40 www has CET as default timezone 2014-08-15 09:49:42 ▶ cat DESCRIPTION 2014-08-15 09:49:44 v140423-201-g2996c60 2014-08-15 09:49:53 clandmeter: Does one need to pin both main and testing repo? 2014-08-15 09:50:01 so it adjust the timestamp yes 2014-08-15 09:50:19 zenny: depends which pkgs you need 2014-08-15 09:50:31 but for ovs, testing is fine. 2014-08-15 09:50:40 clandmeter: thanks 2014-08-15 09:51:18 btw is there a rationale somewhere for the switch to musl? 2014-08-15 09:51:45 probably on the ml 2014-08-15 09:53:12 i think musl's maintenance/support is an important one. 2014-08-15 09:54:14 found it, thnx http://lists.alpinelinux.org/alpine-devel/3598.html 2014-08-15 09:54:31 CcxCZ: http://www.etalabs.net/compare_libcs.html 2014-08-15 10:01:56 CcxCZ: check musl release freq, and compare it with uclibc... 2014-08-15 10:05:41 clandmeter: what are the packages that one needs to install kvm? 'apk search kvm' outputs nothing. 2014-08-15 10:07:01 kvm is no package :) 2014-08-15 10:07:11 kvm inside kernel 2014-08-15 10:07:36 well, there used to be a fork of qemu called that 2014-08-15 10:07:42 but it got merged back 2014-08-15 10:07:46 and you need qemu 2014-08-15 10:08:01 apk add qemu-x86_64 2014-08-15 10:08:16 CcxCZ: correct 2014-08-15 10:09:36 zenny: try libvirt, it should make things a bit easier. although im not sure its supports ovs. 2014-08-15 10:13:54 zenny: if you can confirm ovs is working properly, you can request it to be moved to main. In the future we are planning to cleanup testing to reduce workload on unmaintained packages. 2014-08-15 10:15:15 clandmeter: I shall do accordingly. but ovs seems to be outdated at 2.1 version whereas 2.3 has already been released. 2014-08-15 10:15:59 there was somebody who supported it 2014-08-15 10:16:14 but seems he didnt update it for some time (for uknown reasons) 2014-08-15 10:16:19 is there a documentation somewhere for the extra flags apk takes in addition to the commands which are described on the wiki? 2014-08-15 10:19:12 zenny: 2.3.0 build fine on musl, so if you want i can push it. 2014-08-15 10:19:35 CcxCZ: apk help or else src i guess. 2014-08-15 10:22:40 <^7heo_> man apk? 2014-08-15 10:23:00 <^7heo_> apk add apk-doc if it exists :) 2014-08-15 10:25:21 clandmeter: it would be nice. BTW which libvirt package to install libvirt-qemu? 2014-08-15 10:25:29 I don't really see manpage source in the git repo 2014-08-15 10:26:35 clandmeter: I meant ovs 2.3 2014-08-15 10:28:49 where can I find the package signing key? is it on some keyserver? 2014-08-15 10:29:52 alpine-keys 2014-08-15 10:31:11 I'm seeding from non-alpine. is this normal gpg? 2014-08-15 10:34:33 if I just extract the alpine-keys and apk-tools-static to / it will allow pick up the keys properly, right? 2014-08-15 10:36:31 i dont understand your question. 2014-08-15 10:37:11 you want to run apk on non alpine? 2014-08-15 10:37:23 I'm doing the chroot install and I want apk.static to check the keys properly 2014-08-15 10:37:33 either that or a way to check them myself 2014-08-15 10:38:05 if you create an alpine chroot, alpine-keys should be pulled in 2014-08-15 10:38:41 you would only need apk.static outside the chroot 2014-08-15 10:39:44 there's the matter of checking APKINDEX and the initial chroot files 2014-08-15 10:44:11 i guess you can put them in the correct location and it should work 2014-08-15 10:49:13 Is there a guide to working with/setting up ZFS on Alpine? 2014-08-15 10:53:38 is it currently working? there were some issues with musl+zfs 2014-08-15 10:57:29 ohi 2014-08-15 10:57:35 hi 2014-08-15 10:57:47 JT-EC: f15RyWpgGTO6 is working on zfs support 2014-08-15 10:59:02 Oh, cool. Thanks. Just read the update in -devel :) 2014-08-15 11:00:47 f15RyWpgGTO6: ZFS seems to be the only bit missing for a quite good linux-based SmartOS equivalent. 2014-08-15 11:01:09 pls define smartos 2014-08-15 11:01:15 smartos.org 2014-08-15 11:01:38 ok.thx. 2014-08-15 11:01:52 Does docker support ZFS? 2014-08-15 11:02:05 It's a virtualisation OS made by Joyent based on illumos. 2014-08-15 11:02:47 hmmmm, nice. illumos does kvm? 2014-08-15 11:03:05 i had the same question :) 2014-08-15 11:03:21 On x86, yep. No plans for SPARC as Joyent don't have any. 2014-08-15 11:03:59 JT-EC: https://github.com/zfsonlinux/zfs/pull/2380 2014-08-15 11:04:06 It's intel too at the moment, the AMD support is a test patch as Joyent don't have AMD either! 2014-08-15 11:04:27 zenny: yes those are mine. and currently not compiling 2014-08-15 11:04:54 shafire: Docker does not support zfs. It is based on ext4 and aufs 2014-08-15 11:05:05 it supports btrfs 2014-08-15 11:05:15 so thats why I use btrfs too 2014-08-15 11:05:44 afaik docker does not support btrfs either. Maybe things have changed since v 1.0.0 2014-08-15 11:05:58 I am using it :D 2014-08-15 11:06:03 f15RyWpgGTO6: Thanks for confirming. 2014-08-15 11:06:21 docker on bcache btrfs raid1 2014-08-15 11:06:53 shafire: it is aufs on top of btrfs, maybe an extra layer. docker does not work directly on btrfs till v 0.9 2014-08-15 11:07:52 clandmeter: did you push ovs 2.3.0? 2014-08-15 11:08:07 Fri:33 13:07:24 root:alpine~# docker info Containers: 79 Images: 231 Storage Driver: btrfs Execution Driver: native-0.2 Kernel Version: 3.14.16-0-grsec 2014-08-15 11:08:14 does not look like it :S 2014-08-15 11:08:15 zenny: check alpine-devel 2014-08-15 11:08:33 zenny: but i have gotten some kind support, some of the patches have been cherrypicked and are now in zfs. some other patches have gotten some love from some kind contributors. and now i'm trying to merge all this into a new patch that compiles cleanly and works. 2014-08-15 11:09:02 f15RyWpgGTO6: Look forward to. ;-) 2014-08-15 11:09:42 zenny: http://nl.alpinelinux.org/alpine/edge/testing/x86_64/openvswitch-2.3.0-r0.apk 2014-08-15 11:11:41 clandmeter: Thanks. shall pull and install and see how it goes. 2014-08-15 11:13:31 apk add openvswitch@testing 2014-08-15 11:15:24 safire: "The repository tag for world dependency 'openvswitch@testing' " 2014-08-15 11:16:40 safire: "ERROR: Not committing changes due to missing repository tags." 2014-08-15 11:17:01 add @testing before the url in /etc/apk/repositories 2014-08-15 11:17:12 and @edge and normal 2014-08-15 11:20:09 shafire: nope, 'apk add openvswitch@testing' pulled the same old version 2.1. 2014-08-15 11:20:16 ah 2014-08-15 11:20:22 which server do you use? 2014-08-15 11:20:24 wrong repo link 2014-08-15 11:20:31 @edge http://mirror1.hs-esslingen.de/pub/Mirrors/alpine/edge/main 2014-08-15 11:20:32 @testing http://mirror1.hs-esslingen.de/pub/Mirrors/alpine/edge/testing 2014-08-15 11:20:39 zenny: esslingen needs 4h 2014-08-15 11:20:44 use nl.alpinelinux.org 2014-08-15 11:20:48 yes use nl.a.o 2014-08-15 11:21:01 its sitting beside me here ;-) 2014-08-15 11:21:14 thanks. 2014-08-15 11:21:23 i can hear her yawn 2014-08-15 11:21:37 edutel b.v.? clandmeter? 2014-08-15 11:21:41 home server? 2014-08-15 11:22:07 no 2014-08-15 11:22:25 are you working in a rz? 2014-08-15 11:22:38 in a what? 2014-08-15 11:22:54 rechenzentrum 2014-08-15 11:23:05 nein 2014-08-15 11:23:23 in an office with a server room not too far away 2014-08-15 11:24:07 yeah datacenter 2014-08-15 11:24:12 and those pizza boxes make a hell of alot of noise 2014-08-15 11:24:38 i should call our rnd and complain 2014-08-15 11:24:56 but i dont speak chinese :) 2014-08-15 11:27:47 clandmeter: I do, but only one word, Xie Xie. 2014-08-15 11:27:52 ;-) 2014-08-15 11:29:41 in which package can I find gnu ifconfig 2014-08-15 11:30:52 net-tools 2014-08-15 11:31:00 according to http://packages.ubuntu.com 2014-08-15 11:34:37 what is the ideal way to configure bridge in alpine? setup-interfaces or directly making changes to /etc/network/interfaces or ovs-vsctl add-br br0? 2014-08-15 11:35:25 clandmeter: thanks, ipv6 is working now :) 2014-08-15 11:38:36 clandmeter: does openvswitch-2.3.0 have init script at /etc/init.d/openvswitch-switch? Trying to add a bridge gives an error "ovs-vsctl: unix:/var/run/openvswitch/db.sock: database connection failed (No such file or directory)" meaning openvswitch is not started. 2014-08-15 11:38:54 Already loaded the module: 2014-08-15 11:39:07 # lsmod | grep openvswitch 2014-08-15 11:39:08 openvswitch 52313 0 2014-08-15 11:39:08 libcrc32c 1188 1 openvswitch 2014-08-15 11:39:08 vxlan 25364 1 openvswitch 2014-08-15 11:41:05 Okay, found it. it is called /etc/init.d/ovs-vswitchd in alpine. 2014-08-15 12:01:57 zenny: i dont know anything about ovs, but i saw there are 2 init scripts. 2014-08-15 12:39:53 hmpf 2014-08-15 12:39:55 GDBus.Error:org.freedesktop.DBus.Error.UnknownMethod: Method "Activate" with signature "" on interface "org.freedesktop.ConsoleKit.Session" doesn't exist 2014-08-15 12:40:04 that happens when trying to run lightdm 2014-08-15 12:44:29 you're using wrong init system apparently 2014-08-15 12:44:44 my bad 2014-08-15 12:44:45 on that note, any reason why there's no runit package? 2014-08-15 12:44:45 it runs 2014-08-15 12:45:02 it was /var/run/lightdm that had bad perms 2014-08-15 12:45:15 CcxCZ: we have runit in busybox i think 2014-08-15 12:45:26 oh, its not enabled 2014-08-15 12:45:39 huh? 2014-08-15 12:45:55 i think busybox provides a runit 2014-08-15 12:46:03 as in http://smarden.org/runit/ ? 2014-08-15 12:46:14 iirc, yes it does 2014-08-15 12:47:03 hum 2014-08-15 12:47:07 seems like im wrong 2014-08-15 12:47:25 it provides runsv, runsvdir svlogd etc 2014-08-15 12:47:29 but not runit 2014-08-15 12:47:45 yup i think we want a runit package 2014-08-15 12:50:37 or one woulc just run runsvdir from bb init 2014-08-15 12:50:45 s/wouldc/could/ 2014-08-15 12:51:36 but yeah, a package of the upstream version would be nice if someone wan't to run runit as pid 1 2014-08-15 12:51:45 i'd like to run runit as pid 1 + something else as service manager, which does 'push' instead of 'pull' 2014-08-15 12:52:04 hmm, I currently use runit-init only in vservers, sysvinit+runsv on the main host (both gentoo) 2014-08-15 12:52:30 would like to migrate bulk of it to alpine where sensible 2014-08-15 12:52:49 what do you mean runit does 'pull'? 2014-08-15 12:53:08 it gets SIGCHLD when the child dies 2014-08-15 12:53:21 you ask for services and wait for the hardware to show up 2014-08-15 12:53:31 rather than init service on hardware detection 2014-08-15 12:54:08 I have scripts that enable services based on preconditions (including hw) 2014-08-15 12:54:20 i would also like and contribute to runit 2014-08-15 12:54:25 or you just start the service before hardware is available 2014-08-15 12:54:35 and it dies, but is restarted 2014-08-15 12:54:49 i'd like trigger network start once the depended hw shows up 2014-08-15 12:54:52 eg eth0 2014-08-15 12:55:01 uggedal: yeah, that's what Gerrit suggests 2014-08-15 12:55:28 I think I like the simpler model of not using dependencies 2014-08-15 12:55:35 ncopa: yes, I have rfkill -> wpa_supplicant -> dhcpcd set up like that 2014-08-15 12:56:04 uggedal: i need to think of it 2014-08-15 12:56:11 i'd like to avoid wait 2014-08-15 12:56:14 at least as a start 2014-08-15 12:56:48 well, by not using deps you don't wait, just retry 2014-08-15 12:57:11 you need wait slightly before you retry 2014-08-15 12:57:20 or cpu goes to 100% 2014-08-15 12:57:42 in any case, i think we want the runit binary 2014-08-15 12:57:57 so people can experiment with it 2014-08-15 12:58:44 runit is actively mantained, there was a release just few days ago, so there might be benefit from using it for runsv* etc. too 2014-08-15 12:59:38 there are some upcoming patches about one funny race condition too 2014-08-15 13:01:24 btw any experiments with toybox yet? 2014-08-15 13:02:15 http://landley.net/code/toybox/ 2014-08-15 13:02:59 nope 2014-08-15 13:03:10 but i have heard about toybox for a while 2014-08-15 13:06:00 I wonder what's the difference between the busybox one and Pape's last release 2014-08-15 13:06:35 gerrit, i believe pape is his familyname 2014-08-15 13:06:38 we use /var/service as defautl service dir? 2014-08-15 13:06:43 or /run/service 2014-08-15 13:06:57 or just keep the default /service 2014-08-15 13:07:21 i assume that /var might need to get mounted 2014-08-15 13:07:58 we install the tools in /sbin or in /usr/bin ? 2014-08-15 13:07:59 dunno, I use /run/service which is on tmpfs on gentoo 2014-08-15 13:08:14 thats why i suggest /run 2014-08-15 13:08:17 due to its tmpfs 2014-08-15 13:08:33 there's ongoing effort to integrate it into openrc 2014-08-15 13:10:50 well, with, not into 2014-08-15 13:14:33 void linus uses /var/service -> /var/run/runit/runsvdir/current -> /etc/runit/runsvdir/default 2014-08-15 13:15:38 I do /run/service/foo -> /run/zsv/foo generated from /etc/sv.d/foo 2014-08-15 13:17:04 the one in etc gets sourced basically for all the actions, using shell functions to declare preconditions, checks, how to run, how to shutdown... 2014-08-15 13:17:26 runit pushed 2014-08-15 13:18:42 ? … well you don't want runsvdir watching hardware. but you can make it reload a list of services with simple SIGCONT 2014-08-15 13:19:16 my approach should not have unnecessary sleeps in it, aside from the ./check loop that sv itself does 2014-08-15 13:21:23 I have listeners for kernel events and set up hooks in several applications (eg. mdev, wpa_cli, ...) that trigger reevaluation of what should be up 2014-08-15 13:22:57 nice 2014-08-15 13:23:07 are these scripts available somewhere? 2014-08-15 13:24:15 http://wpr.cz/ccx/bzr/zsv/ docs are bit out of date unfortunately, but it's what I have running atm on several boxes including my laptop 2014-08-15 13:26:25 nice, thx. will have a look! 2014-08-15 13:28:23 sbin/zsvgen is what generates the services, zsv.exec is entrypoint to everything else 2014-08-15 13:31:23 you might like the svs script here also http://wpr.cz/ccx/bzr/usv/ 2014-08-15 13:45:30 one last thing I seem to need to migrate the host to alpine would be to convert my syslog-ng -> multiple multilogs setup to rsyslog -> multiple svlogd … or is there anything better matching my current setup in stock alpine? 2014-08-15 13:57:21 perhaps I could use svlogd and a bunch of named pipes too 2014-08-15 14:53:38 CcxCZ: what is the benefit of using zsh? are there particular features that are unique for service scripts? 2014-08-15 14:58:55 f15RyWpgGTO6: it's the easiest to prototype for me, but indeed, I use the function autoloading and $fpath extensively 2014-08-15 14:59:31 and potentialy using zcompile to pre-parse all of the scripts into one file for some speed benefit 2014-08-15 15:07:04 f15RyWpgGTO6: it's a prototype still, I mainly wanted to discover the features that I need and get it working. I'm not adverse to posix sh rewrite, but this makes extending it way easier imo 2014-08-15 17:20:34 What is the right way to load modules and init scripts at boot time in alpine permanently? 2014-08-15 22:14:41 So Alpine is not meant to be installed? 2014-08-15 22:16:13 on the contrary 2014-08-15 22:17:07 That is what I read from the Wiki 2014-08-15 22:22:39 ok so is apk like apt or more like nixos's nix package manager? 2014-08-15 22:23:58 more like a faster apt 2014-08-15 22:24:56 for traditional disk installs, see: http://wiki.alpinelinux.org/wiki/Install_to_disk 2014-08-15 22:30:05 do you have rootless X in your packages 2014-08-15 22:58:32 all my alpine are full installs... 2014-08-16 01:32:06 hi: grsec denies resource overstep and causes segmentation fault to openvswitch in alpine-edge? Any idea? 2014-08-16 11:49:59 somebody is trying to get nvidia gl working with musl http://pastebin.com/raw.php?i=cMiv8UVp 2014-08-16 11:50:03 he seems pretty close 2014-08-16 11:52:19 That's cool 2014-08-16 11:54:53 $ man minitube 2014-08-16 11:54:53 No manual entry for minitube 2014-08-16 11:54:58 Why do people keep doing this 2014-08-16 11:55:15 ...and then openfbtd remembered his scripts don't have mans too 2014-08-16 11:56:01 did you install -doc package? 2014-08-16 11:56:26 if its available... 2014-08-16 12:00:26 It's not on alpine. 2014-08-16 12:01:40 Yeah, minitube just doesn't have a man. At all. 2014-08-16 12:02:10 To be fair, not that it needs to. 2014-08-16 12:02:32 It has four flags including --help 2014-08-16 12:02:54 i didnt even know what it was untill now i googled it 2014-08-16 16:07:16 Is alpine vuln to remote busybox exploits? 2014-08-16 16:07:30 Or is the version greater than i think 1.4 or whatever that vuln version is 2014-08-16 16:21:57 You can look 2014-08-16 16:22:29 https://git.fleshless.org/aports/tree/main/busybox/APKBUILD 2014-08-16 16:45:07 oh cool, 1.22.1 2014-08-16 17:02:07 Howdy; I'm trying to boot 3.0.3 off a USB disk, and after after the Syslinux prompt, I get a bunch of error messages about musl, busybox, alpine-baselayout, &c exiting with error 1, and a few "no such file or directory" errors. 2014-08-16 17:02:27 I made the USB disk by following the "manual" method. It's using vfat and I checked to make sure 'usbdisk' was specified in syslinux.cfg 2014-08-16 17:03:54 usererror, which remote busybox issues? 2014-08-16 17:04:37 there was an exploit I had to patch for a NAS device when I saw some interesting traffic 2014-08-16 17:04:49 was busybox related, probably a month ago 2014-08-16 17:04:59 i think it was for an older version, good to see you guys are on 1.22.1 2014-08-16 17:22:03 Alright, I tried it again on a USB drive I didn't get for free from my university and it worked fine. 2014-08-16 17:36:15 Err, however, apk won't work; any apk command spits back: 2014-08-16 17:36:27 WARNING: Ignoring APKINDEX.b83d5251.tar.gz: No such file or directory. 2014-08-16 17:43:21 Okay, wow, it's like I forget how to use apk. Ignore me. 2014-08-17 17:59:57 Does anyone know if the raspi Maynard de or whatever has a package for alpine? 2014-08-18 07:32:46 hi , on afresh 3.0.3 x64 stable install (thinkpad x200) intelsound is loaded completely but vlc doesnt show any devices ,alsamixer shows them.-> the vlc-plugins (vlc-plugins-audio_output)are missing on the mirror .where do i find them? thanx 2014-08-18 07:55:12 <^7heo_> moin 2014-08-18 20:30:26 Does alpine support ARM? 2014-08-18 20:30:47 Would be cool to have an ARMv7 FS i could add to my board 2014-08-18 20:39:13 UserError: Yes, it is possible to build alpine for arm. 2014-08-18 20:39:44 Google alpine raspberry pi for details. Not officially supported, no repos OTOH. 2014-08-18 20:39:52 rats 2014-08-18 20:42:05 we have repo 2014-08-18 20:42:12 ARMv7+ ? 2014-08-18 20:43:15 only armhf 2014-08-18 20:43:27 right yes 2014-08-18 20:43:30 http://nl.alpinelinux.org/alpine/edge/main/armhf 2014-08-18 20:43:46 and we have an image for rpi 2014-08-18 20:45:32 http://dev.alpinelinux.org/~tteras/alpine-edge-140529-armhf.rpi.tar.gz 2014-08-18 20:48:38 Is the rpi image different in any way 2014-08-18 20:48:46 from the regular repos 2014-08-18 20:48:54 other than just having kernel and prob workaround packages etc 2014-08-18 20:49:08 Amazing, gtk3, awesome awesome 2014-08-18 20:50:31 Hi, sorry for my bad connection ... on my standard x64stable install 3.0.3, i want to use qemu/kvm.it doesnt accept standard syntax qemu-i386 -hda -m -enable-kvm and shows me options about offsets etc i never saw before .is that a different version ? how do i use qemu with kvm under alpine as host? thanx 2014-08-18 20:54:40 (...kvm_intel is loaded) 2014-08-18 20:54:54 olfylyj: shouldnt you just use qemu-x86_64? 2014-08-18 20:55:54 @clandmeter: thanx but till now these relate to the guest architecture not the host 2014-08-18 20:56:52 i386 guest on x64 host needs qemui386 2014-08-18 20:57:32 at least on debian 2014-08-18 20:57:46 and freebsd 2014-08-18 20:57:55 you could try :) 2014-08-18 20:57:57 its free 2014-08-18 20:58:36 i remember somebody mention it. so dont blame me. 2014-08-18 21:00:15 :) i thought maybe its aproblen with the grsec extensions ... 2014-08-18 21:00:32 ...a problem , sorry 2014-08-18 21:02:55 part of our infra runs on grsec and kvm 2014-08-18 21:03:10 dmesg should tell you if it kicks in 2014-08-18 21:04:55 so in short qemu -hda /blah -m 2222 -enable-kvm shouuuuld work? (/dev/kvm is chmod a+rww) 2014-08-18 21:05:48 i dont know, i mostly use libvirt to manage it. 2014-08-18 21:06:05 but i can confirm it works on 2.7 (and probalby 3.0 too) 2014-08-18 21:08:00 hmm,thanks would you have alink to a alpine-compatible libvirt-tutorial/doku? 2014-08-18 21:09:53 olfylyj: sorry, i dont think there is anything on the wiki 2014-08-18 21:11:03 you can try other dists wiki's for instance https://wiki.archlinux.org/index.php/libvirt 2014-08-18 21:13:46 yep thanx, archwiki is nice , i just was suspicious on the influence of grsec ... i go and try that 2014-08-19 14:30:23 anyone have hvm domUs working on 3.x now? 2014-08-19 14:36:08 <^7heo> (feedback so you know that someone read that: not me) 2014-08-19 14:36:22 <^7heo> (I used to, but my server burnt down) 2014-08-19 14:43:04 burnt down!?! :> 2014-08-19 14:44:34 <^7heo> yeah 2014-08-19 14:44:37 <^7heo> I still have to determine the cause 2014-08-19 14:44:43 <^7heo> but I don't have any way to do so for now., 2014-08-19 14:44:44 <^7heo> so... 2014-08-19 14:45:43 sorry to hear :( 2014-08-19 14:45:48 that's pretty horribly 2014-08-19 14:45:50 *e 2014-08-19 14:47:03 <^7heo> yeah 2014-08-19 14:47:06 <^7heo> but what can I do... 2014-08-19 14:47:34 get a new one :) 2014-08-19 15:16:05 <^7heo> mlaine: I have no salary yet. 2014-08-19 15:16:12 <^7heo> mlaine: I have to wait 'til the end of the month 2014-08-19 15:16:20 <^7heo> and then I have a freaking impressive list of things to get. 2014-08-19 15:16:29 <^7heo> from a bed to some ram for my laptop 2014-08-19 15:16:31 <^7heo> etc. 2014-08-19 15:16:41 <^7heo> so thanks for the suggestion, but I can't right now :) 2014-08-19 15:17:06 no worries 2014-08-19 15:33:53 ^7heo: switched jobs? the first month can stretch pretty long hehe 2014-08-19 15:39:24 <^7heo> darkfader: yeah 2014-08-19 15:39:42 <^7heo> actually, I got fired from the previous one, for knowing the ex-gf or my coworker. 2014-08-19 15:41:22 well, sounds like it's better to know her and not work there. if they act like that. 2014-08-19 15:42:03 the person who trained me here did the opposite & told me she'd not even have talked to me if she'd known my ex-employer 2014-08-19 15:42:22 people are quick with wrong judgement? :) 2014-08-19 16:38:31 <^7heo> inkscape is awesome 2014-08-19 18:05:00 Hi there... can someone help me with awall to allow sip without connection tracking? 2014-08-19 18:05:11 I' tried to use the following: 2014-08-19 18:05:12 "filter": [ 2014-08-19 18:05:12 { 2014-08-19 18:05:12 "out": "_fw", 2014-08-19 18:05:13 "service": [ "sip", "sip-tls" ], 2014-08-19 18:05:15 "action": "accept", 2014-08-19 18:05:17 "no-track": true 2014-08-19 18:05:19 }, 2014-08-19 18:05:21 { 2014-08-19 18:05:23 "in": "_fw", 2014-08-19 18:05:25 "service": [ "sip", "sip-tls" ], 2014-08-19 18:05:27 "action": "accept", 2014-08-19 18:05:29 "no-track": true 2014-08-19 18:05:31 } 2014-08-19 18:05:33 ] 2014-08-19 18:05:35 } 2014-08-19 18:05:37 but when I try to activate the following error is printed: 2014-08-19 18:05:39 awall: filter 3 (server-sip): Tracking required by service 2014-08-19 18:05:49 in both AL 3.0.3 and 2.7.9 2014-08-19 18:06:56 Looks that it's because the sip service definition in /usr/share/awall/mandatory/services.json includes "ct-helper": "sip" 2014-08-19 18:07:04 "sip": [ 2014-08-19 18:07:04 { "proto": "udp", "port": 5060, "ct-helper": "sip" }, 2014-08-19 18:07:04 { "proto": "tcp", "port": 5060, "ct-helper": "sip" } 2014-08-19 18:07:04 ] 2014-08-19 18:07:12 there is anyway to bypass that? 2014-08-19 18:12:56 I expect that's because SIP uses other ports to carry the media streams so needs some tracking to know which ports to open in conjunction with the SIP. I assume you could try bypassing it if awall lets you define other services. Just define a sipuntracked or something with the same definition but without the helper. 2014-08-19 18:13:09 I haven't used awall though so that may not work. 2014-08-19 18:17:16 Meths, will try that way.. 2014-08-19 18:29:21 Meths, this will bypass the error... tks. 2014-08-19 19:17:08 I'm thinking about using Alpine as the base OS for a scientific cluster... however the guys from this area are crazy about performance... Fortran is dominant and they want to lose as little as possible with overhead... do you think Alpine fits? 2014-08-19 19:22:23 haole_: its a-ok as long you use your kernel 2014-08-19 19:22:28 you have to in scientific computin anyway 2014-08-19 19:25:04 katlogic: why is that? what's enabled by default that might hurt me? 2014-08-19 19:26:25 grsec overhead is small, but present. also hinders some troubleshooting because it overzealously censors kernel. for example try cat /proc/$$/maps 2014-08-19 19:26:37 figuring out how to disable all that crap ... its just better to not run grsec at all 2014-08-19 19:28:06 hmm, maybe just rebuild some core pieces with the intel compiler if they are performace-hungry 2014-08-19 19:28:22 a eval version should be free 2014-08-19 19:28:31 and i remember something like 20% difference for postgres 2014-08-19 19:28:49 no idea about fortran ;) 2014-08-19 19:29:05 no need for icc, might hurt actually 2014-08-19 19:29:28 musl is well otpimized for gcc, naturally things like fortran libs is icc domain :) 2014-08-19 19:29:37 the difference in some aspects are huge, eg. malloc speed glibc vs musl 2014-08-19 19:29:55 katlogic: thanks for correcting me then, i didn't know half of that 2014-08-19 19:30:43 darkfader: you might also well run into some trouble, because alpine is still pretty much niche. depending on your app really. 2014-08-19 19:30:53 just fortran stuff, or random assortment of c/c++ too? 2014-08-19 19:31:01 hi 2014-08-19 19:31:29 katlogic: ??? 2014-08-19 19:31:49 katlogic: fortran applications do not really use the libc for maths 2014-08-19 19:31:55 katlogic: they use blas 2014-08-19 19:31:55 kaniini: i know 2014-08-19 19:32:07 katlogic: musl should have no bearing either way on fortran performance 2014-08-19 19:32:17 kaniini: but usually things live in bigger ponds than just pure fortran. 2014-08-19 19:32:18 kaniini: hi, it's not what he said either :) 2014-08-19 19:33:44 i think that alpine would run perfectly fine on a scientific cluster, considering it's designed to be a lightweight operating system for clusters 2014-08-19 19:35:26 ... untill you get this .tar.gz full of random files which were tested to compile only on redhat 2014-08-19 19:36:47 well 2014-08-19 19:36:51 that would really suck 2014-08-19 19:37:17 scientific. imagine all the stereotypes associated with that. 2014-08-19 19:37:22 those generalizations exist for a reason :> 2014-08-19 19:38:09 darkfader: we already use intel's compilers for what we can - we have the license 2014-08-19 19:38:54 I'm between Alpine and TinyCore with basically the same concerns 2014-08-19 19:39:16 from what I can see, Alpine tries to do more for me but is more complex in some regard 2014-08-19 19:39:16 katlogic: i'm running megacli through a debian chroot for a year now ;) 2014-08-19 19:39:21 but I've barely scratched both systems 2014-08-19 19:39:23 haole_: tinycore sucks, way too stripped down imo :> 2014-08-19 19:39:45 alpine actually feels like a distro, not just floppy drive live emergency shell. 2014-08-19 19:39:52 haole_: keep testing, imho alpine is the most unix-like linux distro you can get 2014-08-19 19:40:13 yeah, TinyCore feels alien but there is coolness in that haha 2014-08-19 19:40:14 as in, it won't step on it's toes at all 2014-08-19 19:40:49 darkfader: scary :/ might as well just prepare out the libs it depends on and package it with LD_LIBRARY_PATH wrapper or something :) 2014-08-19 19:40:53 ugly necessary /opt evil 2014-08-19 19:41:08 katlogic: it works far too well now 2014-08-19 19:48:00 try running megacli on alpine 3 2014-08-19 19:49:28 /lib/ld-musl-x86_64.so.1 /path/to/megacli 2014-08-19 19:49:31 it should work 2014-08-19 19:54:25 haole_: If you want to go crazy with performance tuning then probably gentoo could be the way. I think last fosdem there was some talk about performance benchmarks of various build flags using gentoo, BLAS implementations were the main reason for it. There are experimental uclibc/musl versions. I'd probably try some vserver-based setup (I don't think there's significant performance hit for vserver) with 2014-08-19 19:54:27 alpine host and you can try out various guests 2014-08-19 20:13:45 CcxCZ: we don't need to fine tune every binary of the server, only the ones for the simulation software, which are already compiled with the craziest optimizations possible 2014-08-19 20:13:52 I custom kernel might give us some benefit as well 2014-08-19 20:14:21 and I intend to use docker to deploy the systems, so I want a minimal and straightforward linux to avoid having to deal with unnecessary complexity 2014-08-19 20:16:42 systems == cluster nodes 2014-08-19 20:17:20 I'm working with low budgets clusters, so they add any computer that they can get their hands on 2014-08-19 20:17:59 I've first tried a setup with automatic OS installation (debian + FAI), so any computer plugged in the network that was configured to be booted from LAN was going to be formatted and configured to be used by the cluster 2014-08-19 20:18:16 however, the crews didn't have the skills to maintain such a system, like adding extra software :( 2014-08-19 20:18:37 so I'm falling back to an even simpler approach: install whatever piece of junk you want in the system, as long as it run docker 2014-08-19 20:18:43 and the copy over the files! I hope it works :) 2014-08-19 21:08:43 sounds good I guess. I have no experience with docker personally. Thing is, you may want to have different optimizations for different boxes, since you seem to have heterogenous hardware. …hrm, can't find the benchmark suite I mentioned 2014-08-19 22:35:13 so alpine keeps the /bin /usr/bin separation? 2014-08-19 22:35:36 or is /usr/bin just for compat? 2014-08-20 05:25:54 Yes, /bin and /usr/bin are separate directories on Alpine. 2014-08-20 05:29:33 ok 2014-08-20 05:29:40 I guess all those busybox symlinks confused me 2014-08-20 08:53:03 Good Morning all! 2014-08-20 08:55:07 mornings 2014-08-20 09:42:40 <^7heo_> moin 2014-08-20 19:09:59 hey 2014-08-20 19:10:02 any news on the ARM port of Alpine? 2014-08-20 19:10:08 my friend is really interested in running Alpine on an ARM device 2014-08-20 20:01:14 There is a raspi image somewhere 2014-08-20 20:01:27 errrr hold up i found the repo 2014-08-20 20:01:39 http://nl.alpinelinux.org/alpine/edge/main/armhf/ 2014-08-20 20:18:23 neat :) 2014-08-20 20:18:40 bin repos 2014-08-20 20:20:32 there is a somewhat outdated edge rootfs for rpi here: http://dev.alpinelinux.org/~tteras/alpine-edge-140529-armhf.rpi.tar.gz 2014-08-20 21:23:21 Awesome, thanks for the location I was going to make a python script to build userland for use in android chroot 2014-08-20 21:23:26 but lost the location 2014-08-21 07:42:39 ping darkfader :) 2014-08-21 07:48:18 darkfader , I searched out that you ever chat about this : Kernel BUG at drivers/xen/blkfront/blkfront.c:595 during pvmove , just want to check is this fixed ? I meet this recently too 2014-08-21 07:49:17 the message log is about 1 year ago ... hope you still can recall anything, ping me back please if you saw this message. thanks! 2014-08-21 08:16:52 <^7heo_> moin @ 2014-08-21 09:17:54 Hi, pygrub don't find kernel in DomU (alpine-xen 3.0.3, alpine pv 2.7.9) why? 2014-08-21 09:24:03 <^7heo_> because pygrub sucks. 2014-08-21 09:24:15 <^7heo_> Please use pvgrup. 2014-08-21 09:24:31 <^7heo_> pvgrub even 2014-08-21 09:24:41 ok will probe 2014-08-21 09:24:44 <^7heo_> danke. 2014-08-21 09:24:59 deutsch gehts auch 2014-08-21 09:25:04 <^7heo_> :P 2014-08-21 09:25:19 <^7heo_> ja, aber es ist ein Englisch channel hier :P 2014-08-21 09:25:43 Richtig! 2014-08-21 09:26:44 er det nå vi skal begynne snakke norsk? 2014-08-21 09:26:45 (that's almost all of the German I know) 2014-08-21 09:27:05 <^7heo_> ncopa: sorry, I don't understand swedish :P 2014-08-21 09:27:27 <^7heo_> or whatever you use in norway :P 2014-08-21 09:27:32 !! 2014-08-21 09:27:41 ha! han tror jag pratade svenska! 2014-08-21 09:27:47 :D 2014-08-21 09:27:47 <^7heo_> :P 2014-08-21 09:30:18 pvgrub ?? unbekannt 2014-08-21 09:30:55 <^7heo_> http://wiki.xen.org/wiki/PvGrub 2014-08-21 09:31:04 <^7heo_> I think I did write most of the doc' there :P 2014-08-21 09:31:37 <^7heo_> and yes you might have to build a couple of stuff. 2014-08-21 09:33:58 ok. howto build with alpine? 2014-08-21 09:34:27 <^7heo_> I've never done that, but it shouldn't be very different 2014-08-21 09:34:28 <^7heo_> git clone 2014-08-21 09:34:30 <^7heo_> then configure 2014-08-21 09:34:31 <^7heo_> then make 2014-08-21 09:34:37 <^7heo_> then check any compilation problem 2014-08-21 09:34:50 <^7heo_> I'm sorry I can't provide you with the pvgrub binary right now, but I'm at work 2014-08-21 09:34:53 <^7heo_> and I have stuff to do. 2014-08-21 09:35:08 <^7heo_> however, if you can afford to wait a bit, I might be able to do it for you. 2014-08-21 09:35:18 <^7heo_> oh, what the heck 2014-08-21 09:35:41 <^7heo_> I can compile that at work, we don't have i7 4770K CPUs for no reason.... 2014-08-21 09:35:57 <^7heo_> I tell you when it's finished ;) 2014-08-21 09:36:01 Thank you! I'm waiting. 2014-08-21 09:38:43 <^7heo_> it's also a good thing that we're using VDSL here. 2014-08-21 09:38:49 <^7heo_> cloning is fast :P 2014-08-21 09:41:35 <^7heo_> I restarted make with the -j8 option 2014-08-21 09:41:41 <^7heo_> it wasn't using all my CPUs 2014-08-21 09:42:05 <^7heo_> and then it's suddently super fast. 2014-08-21 09:43:03 Gib ALLLES! 2014-08-21 09:43:06 <^7heo_> brb, configuring a router. 2014-08-21 09:47:29 <^7heo_> damnit 2014-08-21 09:47:34 <^7heo_> computer froze. 2014-08-21 09:47:38 Hello again 2014-08-21 09:47:59 it was to much 2014-08-21 09:49:04 <^7heo_> nah 2014-08-21 09:49:08 <^7heo_> I think it's more of a bug 2014-08-21 09:49:17 <^7heo_> the computer can take a compilation for sure. 2014-08-21 09:49:45 <^7heo_> I mean, the music is still playing so... 2014-08-21 09:50:26 <^7heo_> -j8 helps a little :P 2014-08-21 09:51:58 Oh Man wöterbuch und Vokabeln suchen. Macht echt keinen Spaß. ;-) 2014-08-21 09:52:01 <^7heo_> it doesn't compile 2014-08-21 09:52:12 <^7heo_> I have to debug why 2014-08-21 09:52:17 <^7heo_> but that's clearly not something I can do now. 2014-08-21 09:52:19 <^7heo_> so... 2014-08-21 09:52:48 <^7heo_> the problem is with the qemu-xen-dir-remote 2014-08-21 09:53:24 <^7heo_> tools/qemu-xen-dir-remote has an empty Makefile... 2014-08-21 09:53:27 <^7heo_> that's why 2014-08-21 09:53:45 and now? 2014-08-21 09:54:24 <^7heo_> now I can checkout an older Makefile 2014-08-21 09:54:27 <^7heo_> but not now. 2014-08-21 09:54:34 <^7heo_> brb, gotta finish my ticket. 2014-08-21 09:55:11 ok thank's 2014-08-21 09:56:20 <^7heo_> but then after, I'll try to fix it. 2014-08-21 09:56:32 today? 2014-08-21 09:56:44 <^7heo_> afterall, if I can find a makefile to build qemu-xen-dir-remote it'll probably go to the next error :P 2014-08-21 09:56:54 <^7heo_> yes, but remember, I'll try :) 2014-08-21 09:57:11 <^7heo_> I'd say that the fastest way to do it if you want to get it is to install a debian somewhere 2014-08-21 09:57:16 :-) 2014-08-21 09:57:17 <^7heo_> and read the doc I linked. 2014-08-21 09:57:29 i will do. 2014-08-21 09:57:30 <^7heo_> because it works great normally. 2014-08-21 09:57:37 <^7heo_> if people didn't commit bullshit it would work :P 2014-08-21 09:58:03 many probs with pygrub 2014-08-21 09:58:22 <^7heo_> yeah it sucks 2014-08-21 09:58:25 <^7heo_> it really does. 2014-08-21 09:58:34 <^7heo_> it's too bad I don't have any pvgrub binary here... 2014-08-21 09:58:40 <^7heo_> because I used to have some. 2014-08-21 09:58:44 <^7heo_> but I'm not working with Xen here. 2014-08-21 09:58:52 <^7heo_> brb, reboot. 2014-08-21 09:58:54 <^7heo_> damn vbox... 2014-08-21 10:00:06 vbox is overhead for 3 DomU (2 Linux 1 W7) on one machine 2014-08-21 10:00:26 <^7heo_> vbox is needed to configure network devices for me, 2014-08-21 10:00:34 ok 2014-08-21 10:02:56 <^7heo_> Debian is broken... 2014-08-21 10:03:05 <^7heo_> they don't have the proper dependencies on packages. 2014-08-21 10:03:06 <^7heo_> that sucks. 2014-08-21 10:03:25 of course it's broken, it's not Alpine! ;¬) 2014-08-21 10:03:29 <^7heo_> yeah 2014-08-21 10:04:16 <^7heo_> brb, reboot again... 2014-08-21 10:10:33 i read the doc. Here is in alpine /usr/lib/xen/boot/pv-grub-x86_32.gz and now? 2014-08-21 10:11:18 <^7heo_> you found the pvgrub gzip? 2014-08-21 10:11:30 ja 2014-08-21 10:20:57 <^7heo_> okay 2014-08-21 10:21:04 <^7heo_> lemme delete my local xen copy then 2014-08-21 10:21:12 <^7heo_> okay so 2014-08-21 10:21:28 ?? 2014-08-21 10:21:43 <^7heo_> place your pv-grub gzip archives somewhere on your Dom0 drive. 2014-08-21 10:22:07 ok 2014-08-21 10:22:26 <^7heo_> and then, in your DomU configuration file, add the two options in http://wiki.xen.org/wiki/PvGrub#Boot_domU_from_local_disk 2014-08-21 10:22:44 <^7heo_> ofc while correcting the path in the kernel option to the right one on your Dom0 disk 2014-08-21 10:23:08 <^7heo_> then, you will have a problem, since your domU drive will not contain a menu.lst file. 2014-08-21 10:23:23 <^7heo_> so you will get grub, and the boot process will stop. 2014-08-21 10:23:27 <^7heo_> tell me when you're there. 2014-08-21 10:23:55 ok i will do so. 2014-08-21 10:27:30 <^7heo_> raah I hate debian 2014-08-21 10:27:43 <^7heo_> it works on pentium 2 computers, but good luck with recent stuff... 2014-08-21 10:30:00 <^7heo_> okay, so I guess that the virtualbox-dkms package builds the vboxdrv kernel module 2014-08-21 10:30:09 <^7heo_> then I need the correct headers... 2014-08-21 10:31:09 <^7heo_> every time I go to type "apt", I end up typing apk, and then correcting... 2014-08-21 10:34:54 <^7heo_> okay, debian is silently failing when it doesn't find the suitable files... 2014-08-21 10:35:03 <^7heo_> piss finished packages exist also in debian... 2014-08-21 10:42:20 <^7heo_> damn 2014-08-21 10:42:28 <^7heo_> and now it fails to build v_v 2014-08-21 10:42:33 <^7heo_> I'm tired of debugging that shit =/ 2014-08-21 10:47:45 <^7heo_> okay, so the problem is to build the fucking vboxdrv kernel object for my kernel. 2014-08-21 10:47:49 <^7heo_> damnit. 2014-08-21 10:47:58 <^7heo_> if only I had some good distribution... 2014-08-21 10:48:04 <^7heo_> but alpine doesn't work with the keyboards here... 2014-08-21 10:48:05 <^7heo_> that sucks 2014-08-21 11:09:08 <^7heo_> after multiple patches 2014-08-21 11:09:13 <^7heo_> it seems to build 2014-08-21 11:11:11 <^7heo_> okay, reboot and check. 2014-08-21 13:07:01 <^7heo_> okay, p7zip makes my system hang. Awesome. 2014-08-21 13:07:07 <^7heo_> So much for debian stability 2014-08-21 13:12:29 xiaowei: I'm surprised that bug is still around 2014-08-21 13:12:41 what happened to you exactly? 2014-08-21 13:12:48 pvmove died on oom? or what? 2014-08-21 13:17:17 <^7heo_> manjaro_user: any news? 2014-08-21 13:18:22 i will create http://wiki.alpinelinux.org/wiki/Create_Alpine_Linux_PV_DomU 2014-08-21 13:18:58 and still hangs on the last boot to use pvgrub 2014-08-21 13:19:27 manjaro_user: on 3.0.x? 2014-08-21 13:19:53 xen 3.0 domu 3.0.3 2014-08-21 13:20:48 boot cren come and then gurb error:15 file not found 2014-08-21 13:20:52 manjaro_user: i dont think that works 2014-08-21 13:21:06 although im no xen guru 2014-08-21 13:21:06 why? 2014-08-21 13:21:16 3x has some ussis with hvmloader 2014-08-21 13:21:19 issues 2014-08-21 13:21:28 better to use 2.7 for now. 2014-08-21 13:21:51 the same procedure? 2014-08-21 13:21:58 i think so 2014-08-21 13:22:05 probably that doc is written for 2.7 2014-08-21 13:22:35 ok on the first is link on 3.0.3 2014-08-21 13:23:21 i think the link for xen should be 2.7 2014-08-21 13:23:31 ncopa: ? 2014-08-21 13:23:38 atleast untill its fixed 2014-08-21 13:23:45 hi? 2014-08-21 13:23:47 i mean for domu is 3.0.3 2014-08-21 13:24:27 manjaro_user: the host is 2.7 or 3.0? 2014-08-21 13:24:37 pv guests should work on alpine 3.0 hosts too 2014-08-21 13:24:40 host 3.0 2014-08-21 13:24:45 its only hvmloader thats broke afaik 2014-08-21 13:24:55 i havent tested myself though 2014-08-21 13:24:56 i dont know the differnece :) 2014-08-21 13:25:18 i just read its broke, whatever it is that doesnt work :) 2014-08-21 13:25:19 pv = paravirtualized. the guest kernels knows it is a virtual guest 2014-08-21 13:25:19 use xen 2.7? 2014-08-21 13:26:00 ncopa: still if half of it is broke, maybe we shouldnt put it on the downloads. 2014-08-21 13:26:01 hvm = harware virtual machine 2014-08-21 13:26:04 i would alwasy use hvm 2014-08-21 13:26:31 i put it there due to i want help with testing and fixing the issue 2014-08-21 13:26:59 i have a dom0 3.0 box here too and i am trying to reproduce the issue 2014-08-21 13:27:10 i kinda think the fix is relatively simple 2014-08-21 13:27:18 i looked at hvmlodaer today 2014-08-21 13:27:56 ah 2014-08-21 13:27:59 i think i did wrong... 2014-08-21 13:29:45 if someone give me step byt step instrsctions how to reproduce the issue 2014-08-21 13:29:49 ...with teaspoon 2014-08-21 13:29:55 then i think i might have a fix soonish 2014-08-21 13:30:08 i did -nostdlib 2014-08-21 13:30:13 but what i meant was -nostdinc 2014-08-21 13:30:16 and now the build fails 2014-08-21 13:30:36 with "no stdint.h" 2014-08-21 13:30:40 i test with alpine-xen 2.7 and domu alpine 2.7 OK? 2014-08-21 13:31:07 manjaro_user: i need simple steps how *I* can reproduce the issue 2014-08-21 13:31:27 i believe alpine-xen 2.7 (dom0) should work 2014-08-21 13:34:24 for xen 3.0 and domu 3.0.3? or other 2014-08-21 13:40:37 <^7heo_> manjaro_user: hangs? 2014-08-21 13:40:40 <^7heo_> what do you mean? 2014-08-21 13:41:30 what shuold i try? which xen and domu? 2014-08-21 13:42:49 manjaro_user: you have to use 2.7 for dom0 2014-08-21 13:42:56 well, for pv you can use 3.x on both 2014-08-21 13:43:02 but hvm currently only works on 2.7 2014-08-21 13:43:22 i have one host running all 3.x (linux only) and another with more hvm stuff that runs 2.7 in dom0 2014-08-21 13:45:09 i try 2.7 first with domu alpine 3.0.3 an then hvm OK? 2014-08-21 13:45:29 yeah, works ok for me 2014-08-21 13:49:15 <^7heo_> hvm is a bit overkill if you're going to boot a Linux 2014-08-21 13:49:19 <^7heo_> but wth... 2014-08-21 13:49:27 manjaro_user: important for 3.0.3 pv domU: install the package compat-grub (i think it's called like that, will check) 2014-08-21 13:49:38 compat-pvgrub 2014-08-21 13:50:06 let me know if you need an example config 2014-08-21 13:50:22 ^7heo_: theory has it that hvm is faster now anyway 2014-08-21 13:50:38 hvm w7 2014-08-21 13:50:50 my theory would be it's faster per vm but still scales like shit as a whole if you look at running 50+ domUs 2014-08-21 13:51:05 but since no dev does that in their laptop or lab esxi it's considered fast :) 2014-08-21 13:51:15 <^7heo_> darkfader: not for the disks // network 2014-08-21 13:51:18 <^7heo_> but well 2014-08-21 13:51:24 hehe 2014-08-21 13:51:28 <^7heo_> and I don't see how hvm would be faster, since it adds overhead 2014-08-21 13:51:47 <^7heo_> okay 2014-08-21 13:51:58 ^7heo_: the cpu page table stuff is apparently faster than the pv hacks 2014-08-21 13:52:04 idk 2014-08-21 13:52:15 <^7heo_> maybe with the vt-x support 2014-08-21 13:52:21 yeah 2014-08-21 13:52:25 <^7heo_> but not on other processors 2014-08-21 13:52:36 <^7heo_> and if you're going to do stuff with intel processors anyway, I'd recommend nova. 2014-08-21 13:52:52 <^7heo_> alpine is much better than debian 2014-08-21 13:52:54 i upgraded a lot of have all those new cpu features... hurt :> 2014-08-21 13:52:58 what's nova? 2014-08-21 13:53:02 <^7heo_> debian running vbox running debian is broken. 2014-08-21 13:53:12 <^7heo_> and I had to spend nearly 5 hours to make vbox work. 2014-08-21 13:53:15 <^7heo_> darkfader: intel nova 2014-08-21 13:53:24 <^7heo_> a microkernel / hypervisor 2014-08-21 13:53:28 <^7heo_> very interesting stuff. 2014-08-21 13:53:30 oh! 2014-08-21 13:53:33 never saw that 2014-08-21 13:53:39 <^7heo_> But as usual with intel, kinda closed and phantom. 2014-08-21 13:53:44 ah ok 2014-08-21 13:53:49 <^7heo_> I mean, it's open source... 2014-08-21 13:53:55 so check back on it in 2 years if it died or got traction 2014-08-21 13:53:55 <^7heo_> but good luck finding doc' 2014-08-21 13:54:02 yeah like "OpenAMT" was :) 2014-08-21 13:54:08 <^7heo_> well, it didn't really move in the last 2-3 years so... 2014-08-21 13:54:23 <^7heo_> I guess that if the tendency stays the same... 2014-08-21 14:33:54 now i have /boot/grub.conf and /boot/grub/menu.list in domu rebot > error:15 file not found 2014-08-21 14:35:00 wasn't it menu.lst? back in grub 0.X… 2014-08-21 14:35:25 ACTION has no experience with anything xen-related 2014-08-21 14:39:45 <^7heo_> manjaro_user: it's normal 2014-08-21 14:39:49 <^7heo_> manjaro_user: I can help you with that. 2014-08-21 14:40:06 <^7heo_> manjaro_user: first thing you have to do is boot, i.e. pass the parameters that should be in the file, manually. 2014-08-21 14:40:23 <^7heo_> basically 2014-08-21 14:40:24 <^7heo_> type 2014-08-21 14:40:28 <^7heo_> root (hd0) 2014-08-21 14:40:39 <^7heo_> kernel /boot/yourkernelname 2014-08-21 14:40:49 <^7heo_> initird /boot/yourinitrd 2014-08-21 14:40:51 <^7heo_> boot 2014-08-21 14:40:56 <^7heo_> that should do it. 2014-08-21 14:41:04 input in grub? 2014-08-21 14:41:05 <^7heo_> and then create a file in your /boot/grub 2014-08-21 14:41:07 <^7heo_> yes 2014-08-21 14:41:11 <^7heo_> in the grub console 2014-08-21 14:41:23 <^7heo_> I'm a little rusty, I've not been doing that for a year 2014-08-21 14:41:24 ok will try 2014-08-21 14:41:37 <^7heo_> you might need to specify the root=/... parameter after your kernel 2014-08-21 14:41:44 <^7heo_> trial && error :) 2014-08-21 14:41:48 <^7heo_> brb, stuff to fix. 2014-08-21 14:42:19 ^7heo_: HVM is faster specially regarding MMU operations, because it uses the hardware-virtualized MMU instead of the PV MMU hypercalls, which are much more slower 2014-08-21 14:43:21 <^7heo_> royger: for that okay 2014-08-21 14:43:27 <^7heo_> but it's not true for everything 2014-08-21 14:43:33 <^7heo_> i.e. I/Os 2014-08-21 14:43:51 ^7heo_: also context switches from user/kernel spaces are much fater on HVM than 64bit PV guests, due to the fact that 64bit PV guests run both kernel and user space on the same proviledge level 2014-08-21 14:44:20 <^7heo_> yes 2014-08-21 14:44:20 ^7heo_: if you use PVHVM it's certainly going to be faster than PV 2014-08-21 14:44:24 <^7heo_> but you can't have hvm on all processors. 2014-08-21 14:44:33 <^7heo_> and yes, PVHVM is another thing. 2014-08-21 14:44:41 <^7heo_> anyway 2014-08-21 14:44:43 <^7heo_> brb, stuff to setup 2014-08-21 14:49:45 root (hd0,0) > filesystemtpye is ext2fs 2014-08-21 14:50:39 root (hd0) Filesystem unkown 2014-08-21 14:52:26 <^7heo_> manjaro_user: well, whatever works :D 2014-08-21 14:52:34 <^7heo_> manjaro_user: might be 0,0 right 2014-08-21 14:53:20 ok, i think i know why xen hvmloader works with uclibc and not with mus libc 2014-08-21 14:53:33 <^7heo_> that's good :P 2014-08-21 14:53:42 question is how to fix it 2014-08-21 14:53:47 yes 0,0 2014-08-21 14:54:22 problem is that uclibc headers will set int64_t differently with gcc -m32 2014-08-21 14:54:37 <^7heo_> manjaro_user: just place the same informatino in your menu.lst 2014-08-21 14:54:37 musl will always set it as the 64 bit host does 2014-08-21 14:55:06 i think one workaround might be to provide its own stdint.h 2014-08-21 14:55:25 and make it use it instead of the host system's 2014-08-21 15:06:19 ok domu 3.0.3 runs :-) 2014-08-21 15:07:15 great :) 2014-08-21 15:07:42 ncopa: if you manage to fix it i'd like to paypal you something for it 2014-08-21 15:08:09 because this is such a horrible issue for me, left me unable to do most things i had planne 2014-08-21 15:08:18 thank's at all 2014-08-21 15:08:31 now hvm w7 2014-08-21 15:17:42 <^7heo_> mwarf 2014-08-21 15:17:51 <^7heo_> don't tell me that you're planning to use IOMMU as well? 2014-08-21 15:21:07 mwarf ? there must only run 1 Aplication permanent nothing else 2014-08-21 15:21:27 <^7heo_> nah but, are you going to do IOMMU? 2014-08-21 15:21:39 yes 2014-08-21 15:21:48 <^7heo_> and the application that you'll have running, isn't it Steam? 2014-08-21 15:21:55 no 2014-08-21 15:22:04 <^7heo_> damn, it's not the same use case as mine... 2014-08-21 15:33:48 ncopa: in which way does uclibc define int64_t? int64_t should always have the same size regardless of the bitness 2014-08-21 15:57:02 got to go, bye for now everyone! 2014-08-21 16:15:07 hvm w7 > domain creation fail 2014-08-21 16:15:25 running in vbox 2014-08-21 16:15:47 manjaro_user: running Xen inside of vbox? 2014-08-21 16:15:54 manjaro_user: i'm not sure if vbox can abstract hvm 2014-08-21 16:16:27 I don't think vbox supports nested VMX, and without that there's no way you will be able to launch HVM domains 2014-08-21 16:16:28 shuold try on realy machine 2014-08-21 16:20:09 ok bye for today 2014-08-21 17:53:14 hi 2014-08-21 17:53:34 how edge is the edge-branch? 2014-08-21 17:55:09 is using edge on a production machine completely out of the question 2014-08-21 18:20:22 mlaine: its up to you 2014-08-21 18:20:39 depends on the usage 2014-08-21 18:27:17 i guess i'll give it a go 2014-08-21 19:12:10 Has anyone experienced timezone issues with Alpine + NodeJS? For some reason, timezones are flipped (e.g. it says I'm +0300 when I should be -0300). 2014-08-21 20:46:41 if this was glibc then a that's known issue with IANA timezones, use Continent/City instead 2014-08-21 20:46:42 dunno how musl handles timezones though 2014-08-22 00:35:21 darkfader: thanks for ping me back:) , yes pvmove died on oom, actually I hit this on a old redhat kernel, after google it I found you ever did something on this problem, so just want to check is there any patch for this ? 2014-08-22 01:09:23 darkfader: I got the oom exactly same line number Kernel BUG at drivers/xen/blkfront/blkfront.c:595 2014-08-22 05:03:22 Hello! 2014-08-22 05:03:22 Sorry to bother you. I need advice. 2014-08-22 05:03:22 I need to traverse a VLAN on the server on the LAN. As I think I just need to make a bridge without an IP address and add into it external and internal VLANs. 2014-08-22 05:03:22 Tell me, please, my assumption is correct? 2014-08-22 05:03:22 I need such a scheme, because the server in the LAN - a voip server asterisk. 2014-08-22 05:15:45 To such a scheme to work, do I need to change any parameters in the kernel? 2014-08-22 08:07:20 Hey there! Is there anybody out there? :) 2014-08-22 08:08:45 let me look out of the window 2014-08-22 08:09:14 just some dog walking around 2014-08-22 08:11:36 Oh, sorry my english. :) Thank you for responding. :) 2014-08-22 08:12:51 clandmeter: Can you give me some advice? About bridge 2 tagged vlans together. 2014-08-22 08:13:18 what do you want to know? 2014-08-22 08:14:17 im not a networking guru :) 2014-08-22 08:14:23 but some here are (atleast better then me) 2014-08-22 08:16:44 If I bridge eth0.566 and eth2.999 in br0 on Alpine Linux, should I use net.bridge.bridge-nf-filter-vlan-tagged = 1 or 0? 2014-08-22 08:21:42 jakal: no idea, maybe ncopa or fabled know this. 2014-08-22 08:22:25 clandmeter: Ok. Thank you for your honest answer. 2014-08-22 09:26:16 xiaowei: i never suspected it to be xen related 2014-08-22 09:26:46 and i think rh had some custom patches for pvmove, especially to avoid it getting stuck 2014-08-22 09:27:01 did you fix it already? 2014-08-22 09:28:05 my story was back in 2009/2010... $customer later moved a few 100 tb with pvmove and didn't have the troubles; i remember that rhel had some newer pvmove than debian then 2014-08-22 09:28:18 or well-patched 2014-08-22 09:29:02 i used pvmove a few times since, worked. but most of the time i just down everything, dd; lvm mirror also works ok for this job and is safer 2014-08-22 09:29:09 since it adds redundancy during the task 2014-08-22 09:29:49 and i don't have 400MB ram dom0s any more due to the error 2014-08-22 09:30:02 putting between 1 and 4GB now 2014-08-22 09:32:06 or (there i'll stop now), it seems most bigger shops tuck in a md layer just in case 2014-08-22 09:41:25 jakal: i read this a few times: bridge-nf-filter-vlan-tagged - pass (1) or don't pass (0) bridged vlan-tagged ARP/IP traffic to arptables/iptables. 2014-08-22 09:43:05 so 0 means it bypassed the firewall, right? but i've also never heard about it before, no idea how you firewall tagged traffic 2014-08-22 09:43:19 in your case it should not apply at all 2014-08-22 09:43:48 since 566 and 999 only have the untagged frames, unless someone injects tags or you do vlan stacking... 2014-08-22 10:07:18 darkfader: Thanks for the answer. 2014-08-22 10:07:18 I combined in the bridge interfaces eth0.566 and eth2.999. The scheme is as follows. 2014-08-22 10:07:18 linux1 with vlan999 <--> switch with vlan 999 <--> alpine linux with bridge and without firewall <--> switch with vlan 566 <--> linux2 with vlan566 2014-08-22 10:07:18 Everywhere traffic is tagged. The idea should work, but traffic between linux and linux2 fails. 2014-08-22 10:12:49 If I remove eth0.566 from bridge, traffic appears between linux1 and alpine. And if I remove eth2.999 from bridge, traffic appears between linux and alpine. 2014-08-22 10:13:06 s/linux/linux2/ 2014-08-23 13:12:04 I have to compile a custom kernel based on the orginal Alpine configuration, but I'm not sure how this works in Alpine 2014-08-23 16:12:23 krjn: cp -r main/linux-grsec yourkernel; cd yourkernel; vi config; vi APKBUILD; abuild -r ? 2014-08-23 17:21:12 uggedal: thanks, i was looking for something equivalent of creating a kernel package on Debian, and this seems to be it 2014-08-24 15:30:35 Hello, can someone help me? During a installation of Alpine Linux on a virtual HD I was unable to select the colemak us variant. Is there any way of installing it after installation? Thanks 2014-08-24 15:32:32 It is a alternative to the querty layout BTW. 2014-08-24 15:34:34 Sorry if this a frequently asked question, I just don't seem to see anything on the wiki about it... 2014-08-24 15:42:10 fraym: setup-keymap? 2014-08-24 15:42:39 Weird 2014-08-24 15:42:46 I get a error while loading keymaps 2014-08-24 15:43:02 cd: line 1: can't cd to //usr/share/bkeymaps 2014-08-24 15:43:04 bug? 2014-08-24 15:43:17 Thanks for helping :) 2014-08-24 15:44:10 apk add bkeymaps? :) 2014-08-24 15:44:38 Just about to type that 2014-08-24 15:44:39 haha 2014-08-24 15:44:42 Thanks 2014-08-24 15:45:03 I have no idea either, never tried i18n with alpine 2014-08-24 15:45:31 WARNING: Ignoring APKINDEX.34910837.tar.gz: No such file or directory 2014-08-24 15:46:02 iirc you should apk update after install 2014-08-24 15:46:06 ERROR: unstatisfiable contrains: bkeymaps (missing): required by world[bkeymaps] 2014-08-24 15:46:07 I did 2014-08-24 15:46:09 Same error 2014-08-24 15:46:12 Really werid 2014-08-24 15:46:20 Weird* 2014-08-24 15:46:35 yeah; the package is listed 2014-08-24 15:46:48 (1/1) Installing bkeymaps (1.13-r3) 2014-08-24 15:47:02 3.0.1 alpine here 2014-08-24 15:47:03 wget: bad address 'alpinelinux.org' 2014-08-24 15:47:07 DNS not right? 2014-08-24 15:47:22 its the nameserver probably 2014-08-24 15:47:23 hehe, alwayas pay attention to earliest error, yeah 2014-08-24 15:47:33 did you try setup-dns? 2014-08-24 15:47:33 get internet working, update apk, install bkeymaps, setup keymap 2014-08-24 15:47:37 not exactly ubuntu huh 2014-08-24 15:48:17 possibly setup-apkrepos too and whatnot 2014-08-24 15:48:18 Now it just reports 404 2014-08-24 15:48:30 yeah, alpinelinux.org sounds like a weird repo 2014-08-24 15:48:34 afaik everyone uses .nl 2014-08-24 15:48:55 setup-interfaces, setup-dns, ifup eth0, setup-apkrepos, apk update 2014-08-24 15:50:31 Default mirror is broken for me 2014-08-24 15:50:33 Changing it 2014-08-24 15:50:37 Thanks for the info 2014-08-24 15:51:23 np 2014-08-24 15:51:28 yeah that's what i usually do :P 2014-08-24 15:51:42 It is working now thanks, it is weird how the default mirror is down 2014-08-24 15:51:50 Haha, thanks man 2014-08-24 15:51:54 :P 2014-08-24 15:53:41 Colemak ain't in setup-keymap, might need a package for it? 2014-08-24 15:54:26 ERROR: unstatisfiable contrains: bkeymaps (missing): required by world[bkeymaps] 2014-08-24 15:54:30 did you find it 2014-08-24 15:54:32 after you fixed dns 2014-08-24 15:54:37 Yes 2014-08-24 15:54:39 oh ok 2014-08-24 15:54:41 It installed fine 2014-08-24 15:54:58 However colemak is not in that package 2014-08-24 15:55:04 oh :( 2014-08-24 15:55:06 ACTION shrugs 2014-08-24 15:55:09 I've never had to mess with keymaps 2014-08-24 15:55:22 Weird how all the BSD's and other linux distro's have it, but alpine linux doesn't 2014-08-24 15:57:49 fraym: guess you need to supply the files then 2014-08-24 15:57:56 Figured so 2014-08-24 15:58:17 the keymap looks extravagant enough, not surprised its not in alpine :) you might want to pullreq it tho 2014-08-24 15:58:53 yeah, I know, it's quite unknown and not so great with languages other than english, but it works for me 2014-08-24 15:59:45 katlogic: How does one pullreq? :) 2014-08-24 16:02:58 this is pre-github. use mailing list :) 2014-08-24 16:03:27 Will look into that, thanks katlogic 2014-08-25 05:00:32 is the alpine edge armhf repo the latest and does python 3 exist there? 2014-08-25 07:12:38 Hi, what can be done when grsec prevents restricts openvswich. See http://imgur.com/fArJkb1? 2014-08-25 07:54:26 Hi, try to start hvm with w7 geust and alpine_xen 3.0.3. ON vncviewer i got black screen. What can i do? 2014-08-25 07:58:00 manjaro_user_: hvmloader is broken on alpine-xen 3.x 2014-08-25 07:58:20 you could copy the hvmloader from a v2.7 package as a workaround 2014-08-25 07:58:29 I am working on the issue though 2014-08-25 07:58:51 when i try 2.7.9 i got mboot.c32 errors 2014-08-25 08:00:06 image 2014-08-25 08:00:29 no, i mean use alpine xen 3.x to boot and everything 2014-08-25 08:00:39 just copy the hvmloader from a v2.7 package 2014-08-25 08:01:25 wget -q -O - http://nl.alpinelinux.org/alpine/v2.7/main/x86_64/xen-4.3.1-r1.apk | tar -zx | usr/lib/xen/boot/hvmloader 2014-08-25 08:01:40 thanks 2014-08-25 08:01:45 that is a temp workaround though 2014-08-25 09:18:28 @ncoap it runs!! Thanks!! 2014-08-25 09:22:57 good! 2014-08-25 09:23:23 i think i might have a proper fix for hvmloader today 2014-08-25 09:24:13 :-) 2014-08-25 09:51:58 manjaro_user_: I think I might have a working fixed hvmloader 2014-08-25 09:52:14 manjaro_user_: do you think you could test it for me? 2014-08-25 09:52:21 ok 2014-08-25 09:52:30 I'll give you an url in a sec 2014-08-25 09:54:50 manjaro_user_: http://dev.alpinelinux.org/~ncopa/xen/hvmloader 2014-08-25 09:56:34 ncopa: oh wow, copying the hvmloader, i should have figured that 2014-08-25 09:56:35 sigh :) 2014-08-25 09:57:07 if someone can confirm that http://dev.alpinelinux.org/~ncopa/xen/hvmloader works i'll push that to edge and 3.0-stable 2014-08-25 09:57:15 its built on a musl host 2014-08-25 09:58:04 ...with some patches I have cooked this morning 2014-08-25 09:59:11 will take day 2014-08-25 09:59:24 lol sorry, too little 2014-08-25 09:59:26 sleep 2014-08-25 10:20:47 ncopa: your fix run here :-) 2014-08-25 10:21:58 what must i do for usb passthrough? 2014-08-25 11:00:19 have no idea :) 2014-08-25 11:00:28 but thanks for testing it 2014-08-25 11:00:46 i'll push it and it will be included in next 3.0.x release 2014-08-25 13:01:05 Is there an apk package creator like for deb 2014-08-25 13:02:20 <^7heo_> http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package3 2014-08-25 13:02:26 <^7heo_> http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package * 2014-08-25 13:02:35 ah, sorry :P 2014-08-25 13:03:28 what are your default compiler settings 2014-08-25 13:04:03 like debian for 32bit is i686 (pentium pro or right before), x64= SSE2 probably AMD something or similar around there on GCC's chart 2014-08-25 13:04:12 <^7heo_> -Wall 2014-08-25 14:05:10 www.cs.tau.ac.il/~tromer/handsoff/ 2014-08-25 14:05:26 I wonder if handshaking with someone for more than 2 sec while typing would be adverse 2014-08-25 15:18:51 Hi there 2014-08-25 15:25:51 <^7heo_> hey 2014-08-25 17:09:18 Is there a way to find out in which package a particular file is? 2014-08-25 17:10:44 <^7heo_> no 2014-08-25 17:10:53 <^7heo_> I tried to implement that, but it's not done yet. 2014-08-25 17:11:10 Ok, thanks :-) 2014-08-25 17:12:25 I was just packaging xorg-util-macros and found out afterwards that the stuff I need is in util-macros. Now I wanted to avoid the same thing for xorg-cf-files :-D 2014-08-25 17:23:29 jomat: only for installed packages. 2014-08-25 17:24:24 Ok, I'll set up a local mirror and index the package contents myself in the meantime 2014-08-25 17:26:17 jomat: you can also check files againt other dists 2014-08-25 17:26:37 i usually use http://packages.ubuntu.com 2014-08-25 17:27:03 Yeah, I did that with archlinux... and that sent me on the wrong path :-) 2014-08-25 17:27:11 <^7heo_> yeah 2014-08-25 17:27:16 :) 2014-08-25 17:27:16 <^7heo_> we need a tool for that 2014-08-25 17:27:22 <^7heo_> like apt-file 2014-08-25 17:27:30 <^7heo_> apk-file maybe? 2014-08-25 17:27:31 <^7heo_> :P 2014-08-25 17:27:35 ubuntu is even more different 2014-08-25 17:28:04 we need somebody to add 2014-08-25 17:28:17 <^7heo_> to add what? 2014-08-25 17:28:18 <^7heo_> the info? 2014-08-25 17:28:31 the feature :) 2014-08-25 17:28:33 <^7heo_> ah yeah 2014-08-25 17:28:40 <^7heo_> I wanted to add it, as aformentionned 2014-08-25 17:28:45 <^7heo_> but I'm really, really tired atm 2014-08-25 17:28:49 <^7heo_> not the best moment for me. 2014-08-25 17:28:53 <^7heo_> however, given some time, maybe. 2014-08-25 17:28:58 <^7heo_> btw, where is the code of apk? 2014-08-25 17:29:09 git.a.o 2014-08-25 17:29:53 <^7heo_> okay 2014-08-25 17:30:09 <^7heo_> http://git.alpinelinux.org/cgit/apk-tools/ That one? 2014-08-25 17:30:18 <^7heo_> seems so 2014-08-25 17:30:20 yes 2014-08-25 17:30:42 <^7heo_> okay, I'll clone that 2014-08-25 17:30:44 <^7heo_> and contribute 2014-08-25 17:31:20 not sure it needs to be part of apk-tools though 2014-08-25 17:31:31 <^7heo_> I can patch apk to do it, or? 2014-08-25 17:31:40 <^7heo_> with a new operation. 2014-08-25 17:31:57 i guess the logic should be on the builders 2014-08-25 17:32:06 <^7heo_> hmm 2014-08-25 17:32:17 <^7heo_> I have to investigate more before I chose what I do anyway 2014-08-25 17:32:24 maybe something insdie buildrepo/abuild 2014-08-25 17:33:28 <^7heo_> btw, I did fork the github repo and patch the kernelconfig files with more common HID drivers as modules... 2014-08-25 17:33:40 <^7heo_> but I have to clean my repo, since I did edit it via the web, it's not only local... 2014-08-25 17:33:46 <^7heo_> I guess I'll just refork 2014-08-25 17:35:01 whatever is handy for you 2014-08-25 17:35:26 ncopa takes core of kernel, so he also takes patches from irc or mail. 2014-08-25 17:35:33 care 2014-08-25 17:37:11 <^7heo_> okay. 2014-08-25 19:02:02 find . -name \*.apk -exec sh -c 'i=`basename $0`;tar -tzf $0|sed s/^/$i:\\//' {} \; >files.txt 2014-08-25 19:02:05 helps me for now 2014-08-25 19:06:32 jomat !!!! <3 2014-08-25 20:46:44 f15RyWpgGTO6: any update about your zfs patch? 2014-08-25 21:07:50 besides there having been lot's of comments unfortunately not from my side, i had my focus diverted by other duties 2014-08-25 21:08:05 but when i catch up with things, i'll pick that up as well 2014-08-26 02:15:45 Hi, just installed alpine stable on my x61s, video and input drivers. Xorg -configure fails selecting the driver, i915 loaded from /etc/modules and nomodeset at boot. However startx does 2014-08-26 02:16:01 but there's no keyboard or mouse input 2014-08-26 02:16:35 then I must reboot for each try \o/ 2014-08-26 02:17:53 If anyone can advice me, i'll be glad ! 2014-08-26 03:23:22 Is gnome 3 in the packages? 2014-08-26 03:24:58 I finally get working X, after upgrading to edge. 2014-08-26 03:25:06 UserError: I dunno 2014-08-26 03:25:38 Just wondering, would like to try and compile the pantheon desktop and package for alpine 2014-08-26 03:26:04 Pantheon ? It's a DE ? 2014-08-26 03:26:07 yes 2014-08-26 03:26:12 http://elementaryos.org 2014-08-26 03:26:36 ah, it's a fork of gnome-shell 2014-08-26 03:26:37 ? 2014-08-26 03:26:45 it uses libmutter 2014-08-26 03:26:49 gala does 2014-08-26 03:27:11 it is written in gtk, is not a fork but if gnome 3 works pretty much pantheon could as far as i know with incompatibilities 2014-08-26 03:27:28 people have it running on arch and debian 2014-08-26 03:27:47 Otherwise i was also going to look at Maynard, which is the new DE for the Raspi 2014-08-26 03:27:52 There is gnome-2.32 2014-08-26 03:27:54 also GTK3 2014-08-26 03:28:10 but i think maynard is wayland 2014-08-26 03:28:12 for* 2014-08-26 03:28:33 I guess gnome after 3.6 will never be packaged due to systemd as dependency 2014-08-26 03:29:20 erggggg 2014-08-26 03:29:31 So maynard is probably a better way to go 2014-08-26 03:29:56 wait but 2014-08-26 03:30:02 is it a gtk3 or libmutter dep 2014-08-26 03:30:10 or just gnomeshell or mutter themselves 2014-08-26 03:30:44 You say it, I never used gnome :/ 2014-08-26 03:31:53 Despite it's an awesome desktop easy to approach environment 2014-08-26 03:36:42 AWESOME http://nl.alpinelinux.org/alpine/edge/main/x86_64/gtk%2b3.0-3.12.2-r0.apk 2014-08-26 05:42:59 When creating packages from git, should we add -git suffix to its name ? 2014-08-26 06:40:19 It's good, I found the answer ! 2014-08-26 06:50:27 In fact, i'm a bit confused. Is this example suitable when the source come only from master.zip ? http://wiki.alpinelinux.org/wiki/APKBUILD_examples:Git_checkout 2014-08-26 07:39:07 Goodbox: abuild has a snapshot function 2014-08-26 07:39:12 abuild --help 2014-08-26 07:46:24 <^7heo_> moin 2014-08-26 07:48:15 clandmeter: ok, trying that :) 2014-08-26 07:48:19 o/ ^7heo_ 2014-08-26 07:50:17 Goodbox: example: http://git.alpinelinux.org/cgit/aports/tree/testing/tvheadend-git/APKBUILD 2014-08-26 07:51:10 thx, then I should create a local repo to set the source ? 2014-08-26 07:51:19 (theorically) 2014-08-26 07:52:00 if you want to share your apkbuild, then the source should be available 2014-08-26 07:52:16 you can replace dev.a.o with your own web location 2014-08-26 07:53:09 it does a git clone, creates the tarbal, set the version number, and upload the tarball wiht rsync to disturl 2014-08-26 07:53:32 you can of course do all of that manuall 2014-08-26 07:54:33 That's right ! I'm making this clear in my mind first. 2014-08-26 07:57:02 git clone should not be done with abuild -r, you should create the tarball and provide it somewhere so if we rebuild it the source is the same again. 2014-08-26 07:58:05 Goodbox: the project you want to build does not have git tags? 2014-08-26 08:00:41 it has well git tags, that's a wm using xcb https://github.com/baskerville/bspwm/tree/0.8.9 2014-08-26 08:01:27 my first package try used the zip, don't slap me :u 2014-08-26 08:02:08 then just use source=$pkgname-$pkgver.tar.gz::https://github.com/baskerville/bspwm/archive/0.8.9.tar.gz 2014-08-26 08:13:42 hmmm okay !I misunderstood the wiki and was gessing the source variable wasn't supporting git at all. To be honest, at this moment I did not know use archive/ver.tar.gz from github. 2014-08-26 08:14:09 then it show my mistake, source supports http 2014-08-26 08:14:43 Thanks you clandmeter :) 2014-08-26 08:15:08 yw 2014-08-26 08:19:26 would it be possible to use something else than busybox in alpine, like sbase + ubase 2014-08-26 08:21:39 you mean side by side or replace busybox? 2014-08-26 08:21:57 replace 2014-08-26 08:24:08 replace wont be that easy 2014-08-26 08:24:30 im not saying its not possible as i dont know sbase/ubase 2014-08-26 08:27:17 abuild automatically adds busybox as dep when any script is detected. 2014-08-26 08:28:05 hmm i see 2014-08-26 08:28:57 ncopa probably knows better how deep busybox is nested into alpine. and what is need to be able to replace it. 2014-08-26 08:29:05 hm 2014-08-26 08:29:44 google tells me uggedal has already tried it. 2014-08-26 08:29:59 http://webcache.googleusercontent.com/search?q=cache:tdzuZyZpcLgJ:https://github.com/uggedal/aports/issues/70+&cd=3&hl=nl&ct=clnk&gl=nl 2014-08-26 08:30:14 i think we should have a provides=/bin/sh or similar 2014-08-26 08:30:43 busybox is hardcoded in abuild atm 2014-08-26 08:31:01 http://git.alpinelinux.org/cgit/abuild/tree/abuild.in#n779 2014-08-26 08:31:26 oh thats for triggers only 2014-08-26 08:31:33 and install scripts 2014-08-26 08:35:44 does busybox suck that much? 2014-08-26 08:36:31 tbh i haven't looked at the code that much 2014-08-26 08:37:52 would like to experiment though 2014-08-26 08:39:16 i think it should work if you patch abuild 2014-08-26 08:40:20 where can i read about sbase/ubse? 2014-08-26 08:40:22 ubase* 2014-08-26 08:40:26 its suckless? 2014-08-26 08:40:30 yes 2014-08-26 08:40:35 http://tools.suckless.org/ 2014-08-26 08:40:49 ncopa: thanks, i might try that 2014-08-26 08:41:35 and in theory you should only need rebuild packages that has trigger or install scripts (eg pre/post-install/upgrade/deinstall) 2014-08-26 08:41:51 if the tools are posix compliant it should work 2014-08-26 08:42:06 it might be that we use some extensions that busybox ash supports 2014-08-26 08:42:19 like local keyword 2014-08-26 08:42:25 im not sure its part of posix 2014-08-26 08:43:01 hm 2014-08-26 08:43:05 sbase 2014-08-26 08:43:12 if you build a package of it 2014-08-26 08:43:15 and install it 2014-08-26 08:43:27 it should in theory just replace busybox version 2014-08-26 08:43:33 just like we do with coreutils 2014-08-26 08:43:42 oh 2014-08-26 08:44:17 only thing the package needs to do is install the binary over the symlink that points to busybox 2014-08-26 08:45:28 one thing though, i doubt those tools will be faster or save space compared to busybox 2014-08-26 08:46:29 busybox has some of the tools as 'nofork' 2014-08-26 08:46:59 using those tools in busybox ash will run significantly faster than if you have to fork/exec a separate binary 2014-08-26 08:56:17 i think i'll try it anyway, thanks for the info 2014-08-26 09:42:59 Since i have bspwm-0.8.4 and its dependency sxhkd-0.5.4 working, should i have to patch makefiles to split the doc ? 2014-08-26 09:43:48 Goodbox: abuild should take care of -doc 2014-08-26 09:44:41 ok, does he declares MANPREFIX ? 2014-08-26 09:45:51 if your package installs docs in the correct location, abuild will move the files to -doc. 2014-08-26 09:46:53 if it doesnt install it, you will have to move it to the correct location in apkbuild package function. 2014-08-26 09:48:10 Great, i'll try sorry for dumbs. 2014-08-26 09:51:03 default_doc() aka doc() (abuild function to move doc files to -doc package) will be run after package() function. 2014-08-26 09:53:22 yeah, i'm a litte longer to start 2014-08-26 14:37:06 hoi 2014-08-26 14:38:53 hey kocka :-) 2014-08-26 14:41:35 hey jomat :) 2014-08-26 14:42:11 was alpine mentioned at camp++?:P 2014-08-26 14:45:20 dunno, probably :-) 2014-08-26 14:49:33 cool 2014-08-26 14:55:40 Confused with data mode installation of 3.0.3 (standard). When the installer asked for the disk, it was 'none', but the installer died without asking for 'request mode' which should be 'data' according to http://wiki.alpinelinux.org/wiki/Installation#Basics, or did I miss something? 2014-08-26 23:27:58 is there a way to determine what caused policykit crap to be installed on alpine, and whether disabling it (chmod -s pkexec) will break anything? 2014-08-27 05:08:23 dalias, "apk del " and it should list what added it but not delete it. you can nuke all top level deps with "apk del -r " but it's easy to shoot yourself in the foot as it'll happily delete everything if you so want 2014-08-27 05:09:49 :-p 2014-08-27 05:10:03 yes i don't want to directly uninstall it but find out how it got pulled in 2014-08-27 05:10:31 it should be possible to install xfce and a useful set of X packages without installing rootkit suid-root 2014-08-27 05:15:45 i don't even know what package pkexec is from 2014-08-27 05:15:51 how do i query that? 2014-08-27 05:18:26 ah info -W 2014-08-27 05:18:31 then search --rdepends 2014-08-27 05:19:04 looks like several power management packages that aren't even working right depend on it, and consolekit depends on it :( 2014-08-27 05:19:21 there should be some way to get pkexec out or make it optional tho 2014-08-27 05:19:32 it's a gaping attack surface 2014-08-27 05:20:47 wtf 2014-08-27 05:20:53 screen on alpine is suid-root? 2014-08-27 05:20:56 that's a gaping hole... 2014-08-27 05:23:15 huh. that doesn't sound good. i'm switched over to tmux, though. 2014-08-27 05:38:31 dalias: file bugs for things like that. if you get shocked by little common issues like mktemps or suid you find shitloads everywhere - that's not to say they don't need to be fixed. it's just to say, there's no "wtf" about it but just file bug, done 2014-08-27 05:51:13 i just did 2014-08-27 06:07:20 morning 2014-08-27 06:07:22 reminds me 2014-08-27 06:07:51 we should patch bb ping to use the datagram icmp ping packets 2014-08-27 06:07:59 to get rid of the suid root ping 2014-08-27 06:08:00 :) 2014-08-27 06:09:03 whatta 2014-08-27 06:09:09 is gnu screen suid root??? 2014-08-27 06:09:40 yes 2014-08-27 06:09:55 stupid defaults in the upstream configure that nobody bothered to check 2014-08-27 06:09:56 btw 2014-08-27 06:10:06 urgh 2014-08-27 06:10:13 yeah must be upstream default 2014-08-27 06:10:19 thanks for reporting it 2014-08-27 06:10:20 a nice improvement to the packaging system would be to disallow creation of suid binaries in packages unless the packager explicitly tags the package that it's allowed to have them 2014-08-27 06:10:27 yes 2014-08-27 06:10:29 good idea 2014-08-27 06:10:39 because lots of upstream crapware does this 2014-08-27 06:10:48 and it's easy to slip through otherwise 2014-08-27 06:10:53 should be relatively easy to implement that 2014-08-27 06:11:36 make abuild exit with error if it finds suid root, unless an options="suid" flag is set or similar 2014-08-27 06:11:43 *nod* 2014-08-27 06:11:44 btw i just msg'd you too 2014-08-27 06:15:10 i also filed several other setuid-related bug reports 2014-08-27 06:16:09 of course getting the options=suid thing into abuild would catch all of these and get them evaluated for whether they're really needed or not 2014-08-27 06:17:16 one other idea to consider, which i kinda mentioned in the intel backlight bug report.. 2014-08-27 06:20:54 would be having wrappers for some or all of these suids which exec a fixed pathname with the real binary 2014-08-27 06:21:12 but only after doing sanity checks in a process with no libs loaded (and possibly static-linked) 2014-08-27 06:22:23 things like checking that the process wasn't started with malicious rlimits, weird fd's inherited, non-whitelisted environment vars, unreasonable command line size (used for spraying), ... 2014-08-27 06:22:40 and only exec'ing the real binary once all the checks pass 2014-08-27 07:00:24 that is an interesting idea indeed 2014-08-27 07:00:57 dalias: feel free to hang around in #alpine-devel too 2014-08-27 07:01:00 ;) 2014-08-27 07:01:47 :) 2014-08-27 07:02:12 i wasn't aware there was a devel channel too. i might but i need to clean up my irc (too many windows! :) 2014-08-27 11:47:08 To be honest i am ashamed that i have to ask, but i can't figure it out. I am unable to mount the current Alpine iso as a loop Device, does anybody have an idea? 2014-08-27 11:47:17 xen01:/xen/iso# mount -o loop alpine-3.0.4-x86_64.iso /mnt 2014-08-27 11:47:18 mount: mounting /dev/loop0 on /mnt failed: Invalid argument 2014-08-27 11:56:46 tze: modprobe loop? 2014-08-27 11:57:57 clandmeter: already loaded, reloading didn't do the trick either 2014-08-27 12:10:37 filesystem 2014-08-27 12:10:55 mount -t iso9660 -o loop alpine-3.0.4-x86_64.iso /mnt 2014-08-27 12:11:23 tze: ^^^ 2014-08-27 12:13:19 Worked, thanks a lot! Could've figured that our by myself ;) 2014-08-27 12:25:40 rer 2014-08-28 02:50:15 Hi, i'm building mumble, the qmake config option "no-bundled-celt" declares wrong INCLUDEPATH=/usr/include/celt in the qmake project file. Is there celt051-dev package, which install celt.h in /usr/includes/celt051/ . What I do is to modify and diff ? 2014-08-28 02:52:36 Goodbox: according to the APKBUILD there should be a celt051-dev package 2014-08-28 02:52:51 just add it to the makedepends 2014-08-28 02:53:28 apk search celt051-dev shows celt051-dev-0.5.1.3-r0 2014-08-28 02:53:40 ACTION is on edge x86_64 if that matters 2014-08-28 02:54:35 Yes it is, however qmake project for mumble declares a different includepath from which celt051-dev installs 2014-08-28 02:56:33 u could patch the qmake project file 2014-08-28 02:56:49 ok, thanks ! 2014-08-28 02:56:53 np 2014-08-28 03:01:03 hmm, isn't strange that the celt header is installed in /usr/include/celt051 ? 2014-08-28 03:02:14 possibly, maybe contact the maintainer 2014-08-28 03:02:25 or see what other distributions are doing 2014-08-28 03:02:43 ok ! 2014-08-28 03:04:25 Goodbox: on archlinux it is also installed in /usr/include/celt051 2014-08-28 03:05:07 ok i'm gonna see the mumble package 2014-08-28 03:05:22 there are two versions of celt apparently 2014-08-28 03:05:34 celt051 is the "spice" version 2014-08-28 03:05:37 i have no idea what that means 2014-08-28 03:05:50 but the normal celt package installes to /usr/include/celt 2014-08-28 03:06:39 ok, i'm running edge, there's just celt-051 2014-08-28 03:07:07 sorry, i was looking at archlinux, they have two celt packages 2014-08-28 03:07:13 https://www.archlinux.org/packages/?q=celt 2014-08-28 03:08:08 all right ! patching seem useful 2014-08-28 03:09:19 I hope mumble get spices on alpine :D 2014-08-28 03:10:18 ive never used mumble before 2014-08-28 03:10:32 but i would love to try it out sometime 2014-08-28 03:10:41 i need to setup alpine on a local machine first though 2014-08-28 03:10:50 The worse thing is dependency for qt for ui and net :/ 2014-08-28 03:10:52 i only have it running on headless servers 2014-08-28 03:11:15 But's a very conference tool 2014-08-28 03:11:20 +nice 2014-08-28 03:13:20 according to wikipedia, mumble defaults to the opus codec now 2014-08-28 03:15:41 yeah, the audio side of this project is pretty good 2014-08-28 03:16:24 speex is really fast 2014-08-28 03:17:49 If I achieve the mumble package i'll try on jack-support patched version 2014-08-28 03:18:22 but i'm really slow at this time 2014-08-28 03:22:02 do you run alpine on desktop/laptop? 2014-08-28 03:22:59 on a x61 and my desktop, never tried on server on as vm 2014-08-28 03:23:42 it feets percfectly to my needs 2014-08-28 03:23:43 how was setting up xorg? 2014-08-28 03:24:02 does mdev handle things like usb drives and such? 2014-08-28 03:24:07 no problem with the dekstop and nouveau 2014-08-28 03:24:20 yeah pretty nice mdev job 2014-08-28 03:24:44 just having to see one thing with intel gpu on the thinkpad 2014-08-28 03:25:11 theorically unset nomodeset and modprobe i915 2014-08-28 03:25:18 oh, ok 2014-08-28 03:25:27 i have a newer intel gpu than the x61 2014-08-28 03:25:35 so im hoping i dont have the same problem 2014-08-28 03:26:20 ugh i said wrong, s/unset/adding/g 2014-08-28 03:27:16 How did you discover alpine ? 2014-08-28 03:28:55 to be honest... the systemd rant on LKML a few weeks ago 2014-08-28 03:29:14 :3 2014-08-28 03:29:23 dunno what LKML is 2014-08-28 03:29:33 linux kernel mailing list 2014-08-28 03:29:48 Huh? Did he mention Alpine in it? 2014-08-28 03:30:00 thx, i'll take a look 2014-08-28 03:30:12 Goodbox, https://lkml.org/lkml/2014/8/12/459 2014-08-28 03:30:24 https://lkml.org/lkml/2014/8/13/623 2014-08-28 03:30:40 Ah 2014-08-28 03:32:38 I'm considering alpine for containers lately. 2014-08-28 03:32:49 At least some of them 2014-08-28 03:34:03 im using it on a VPS and a home server 2014-08-28 03:34:10 hopefully on my laptop soon 2014-08-28 03:34:13 For my part I was running arch when systemd has arrived. I then took few month trying to use a bsd desktop before to switch on alpine. 2014-08-28 03:34:30 lol, that sounds a lot like my journey 2014-08-28 03:34:36 x) 2014-08-28 03:34:39 i used arch for a long time 2014-08-28 03:34:58 last winter i got fed up w/ the direction of arch and used openbsd for a while 2014-08-28 03:35:01 I actually still do, but modified: https://fleshless.org/pages/spark.html 2014-08-28 03:35:30 but openbsd was just too slow and the community is very defensive 2014-08-28 03:35:41 Each bsd where in trouble on different points on my hardware :/ 2014-08-28 03:35:59 in some ways alpine feels like a linux version of a BSD 2014-08-28 03:36:08 yeah the giant kernel lock in the openbsd' smp is really problematic for desktop use 2014-08-28 03:36:16 Arch used to invoke that feeling actually 2014-08-28 03:36:20 ive always been a fan of netbsd and used to run it before i switched to arch 2014-08-28 03:36:21 CRUX still does. 2014-08-28 03:36:27 And Alpine too, of course. 2014-08-28 03:36:34 right, i switched to arch b/c it was the BSD linux w/ rc.conf 2014-08-28 03:36:40 but... then systemd happened 2014-08-28 03:36:52 It was the BSD linux for a bunch of more important reasons :) 2014-08-28 03:36:54 im not totally sold on openrc, but it's been fine 2014-08-28 03:37:01 That's a damn wonderful unix-like :) 2014-08-28 03:37:20 An rc.conf doesn't make you bsd-like! 2014-08-28 03:37:28 true, but it was a selling point 2014-08-28 03:37:36 Internal politics and decisions do. 2014-08-28 03:37:41 Ports do. 2014-08-28 03:38:20 CRUX, Alpine and Arch all still share a wonderful thing btw: publicly available easy to use port systems 2014-08-28 03:38:40 that's where archlinux is very unBSD though 2014-08-28 03:38:44 I think I like Arch's one the best though, just for a bit more flexibility in the package management. 2014-08-28 03:38:51 arch doesnt have a solid base... it changes a lot 2014-08-28 03:38:59 i found netbsd well reactive for desktop, my firefox never get roasted like it, however I'm a bad boy with a nvidia chip :/ 2014-08-28 03:38:59 and very frequently 2014-08-28 03:39:25 CRUX lacks solid dependencies and Alpine lacks a certain feature I quite like: “provides” 2014-08-28 03:39:53 alpine is also is a bit fragile w/ conflicts 2014-08-28 03:40:09 Dummy packages are fine, but I would really prefer dummy packages not being hacky 2014-08-28 03:40:15 like u can install a package that overwrites files belonging to another package 2014-08-28 03:40:25 it will warn u, but if u dont pay attention, u ll miss it 2014-08-28 03:40:45 In the sense that I maybe want openrc-dummy to provide openrc, not have an openrc package with a ridiculously high version number 2014-08-28 03:40:48 That's a hack. 2014-08-28 03:40:52 Not elegant at all 2014-08-28 03:41:29 Goodbox: netbsd has really bad support for modern video cards... im hoping their work on KMS/DRM comes to fruition soon 2014-08-28 03:41:36 (that's from my experience of adapting my packages originally made for Arch to Alpine. And it works, but hacks, hacks everywhere!) 2014-08-28 03:42:08 Also is there no way to tell apk to fuck off and ignore dependencies? 2014-08-28 03:42:22 no idea 2014-08-28 03:42:49 It's annoying as hell that I can't remove openrc from my system without also removing some core packages 2014-08-28 03:43:12 Which is why I had to create a dummy package with a higher version 2014-08-28 03:43:16 what would u replace it with? 2014-08-28 03:43:35 I personally use sinit + spark-rc + watchman 2014-08-28 03:43:43 ^^' wow 2014-08-28 03:43:49 another obscure init scheme 2014-08-28 03:43:53 Yep. 2014-08-28 03:44:06 did you tried s6 ? 2014-08-28 03:44:07 ah, ive actually seen this before 2014-08-28 03:44:08 Two of the above are actually my own projects 2014-08-28 03:44:14 i think when i was looking on the archlinux forums 2014-08-28 03:44:33 pnutzh4x0r, yeah, alad (another Spark user) posted some stuff about Spark on the forums 2014-08-28 03:44:36 And on the wiki 2014-08-28 03:47:14 bwarf why my brain are fucking drawing ombers behind xclients... 2014-08-28 04:06:48 pnutzh4x0r: read. It saddens me to realize how many folks will fall on systemd. Most of users are from ubuntu, debian, fedora, centos or arch. 2014-08-28 04:11:10 ubuntu isnt systemd yet 2014-08-28 04:11:15 they have upstart 2014-08-28 04:11:24 but i think the next LTS will be systemd 2014-08-28 04:11:51 yeah, because the last in date has 2014-08-28 04:12:13 debian has begun the transition and i think that is why we are seeing a resurgence of anti-systemd sentiment 2014-08-28 04:12:23 also centos 7 is the first systemd release for them as well 2014-08-28 04:13:00 that's bad 2014-08-28 04:14:31 did you see http://dontboycottsystemd.org/ ? 2014-08-28 04:14:36 yeah, ive seen that 2014-08-28 04:14:48 haha 2014-08-28 04:15:03 the won is a part of destiny, they dont have to convince anymore 2014-08-28 04:16:08 I dunno remember how fedora folks reacted 2014-08-28 04:16:12 well, in many ways systemd has won 2014-08-28 04:16:24 all the major distros are using it or are switching to it 2014-08-28 04:16:33 slackware and gentoo being the major holdouts 2014-08-28 04:16:41 even those are niche distros 2014-08-28 04:16:47 debian was really the last stand 2014-08-28 04:16:59 uh maybe fedora folks didn't care 2014-08-28 04:17:29 well, lennart works for redhat 2014-08-28 04:17:35 i think fedora got systemd really early 2014-08-28 04:20:07 majority of linux users avoid using good distros 2014-08-28 04:22:10 majority of linux users use android :] 2014-08-28 04:22:30 haha, that's right 2014-08-28 04:28:50 openfbtd: really nice sinit.c 2014-08-28 04:29:07 Goodbox, sinit is the one that is not mine :) 2014-08-28 04:29:10 I just use it 2014-08-28 04:29:19 It's a part of the suckless.org project 2014-08-28 04:30:11 yeah, can you talk me about watchman and spark-rc ? :) 2014-08-28 04:30:21 it's a monitor and a rc script ? 2014-08-28 04:33:35 spark-rc is an rc script, yes, but watchman has kind of a misleading name 2014-08-28 04:34:03 It's actually a service manager in the spirit of both openrc and old arch's init scripts 2014-08-28 04:34:35 nice project ! 2014-08-28 04:34:40 It's basically a central script that sources service configs (that are also scripts) to manage services, both system-wide and user ones 2014-08-28 04:34:51 It does include a respawner though 2014-08-28 04:34:57 cool :) 2014-08-28 04:35:10 which shell is used ? 2014-08-28 04:35:16 bash. Hard dependency. 2014-08-28 04:35:23 ok ! 2014-08-28 04:35:48 Huge ammount of bashisms, won't accept POSIX-ize pull requests. 2014-08-28 04:36:03 Those would mean more external binaries are being used, for one 2014-08-28 04:36:26 watchman is almost fork-less 2014-08-28 04:37:32 yeah, my production uses bash too, despite it is slow and vulnerable, it's powerful 2014-08-28 04:38:00 With scripts like watchman, speed is not a factor 2014-08-28 04:38:16 yep 2014-08-28 04:38:17 Not using external executables is though. 2014-08-28 04:38:28 Forks are expensive and volatile. 2014-08-28 04:40:02 because process creation time can be random ? 2014-08-28 04:40:24 It can, but mostly it's just that forks are bloody slow regardless of that. 2014-08-28 04:40:52 But mostly the problem is that making watchman contained actually ensures portability 2014-08-28 04:40:59 As bash is everywhere 2014-08-28 04:41:08 But coreutils could be different 2014-08-28 04:41:23 I see 2014-08-28 04:42:01 bash didn't work well with busybox ? 2014-08-28 04:42:17 Basically, currently watchman works on most linux AND BSD flavours without little to no effort on my side 2014-08-28 04:42:43 If you have bash, it works. 2014-08-28 04:42:55 Making a script portable with sh is a PITA. 2014-08-28 04:43:04 It's a completely different approach 2014-08-28 04:43:10 One I do not want to adopt. 2014-08-28 04:43:20 :D 2014-08-28 04:43:29 it's only painful if you're used to bash+gnu utils :) 2014-08-28 04:43:39 And even if I did, I don't want to learn all the differences between different coreutils implementations and such 2014-08-28 04:43:43 if you're used to standards, it's not hard to make it portable 2014-08-28 04:43:53 That's the thing, I actively avoid external utils 2014-08-28 04:44:09 standard sh can mostly avoid external utils too 2014-08-28 04:44:26 I've seen big projects using sh. They _have_ to have code that accomodates for different implementations 2014-08-28 04:44:35 no 2014-08-28 04:44:43 they just have to avoid using non-portable constructs 2014-08-28 04:44:46 They _have_ to specifically avoid stuff that works wonky in different sh emulation modes 2014-08-28 04:45:09 As /bin/sh is usually a link 2014-08-28 04:45:13 but that's not very much 2014-08-28 04:45:19 No one actually uses Bourne Shell anymore 2014-08-28 04:46:09 it doesn't matter as long as sh is standards conforming or has minimal tolerable deviations (like bash does) 2014-08-28 04:46:25 And in the end, it's much more comfortable just hard depending on a single language than to willingly inflict pain on yourself by trying to write in limited POSIX shell. 2014-08-28 04:46:41 it's really not that hard 2014-08-28 04:46:45 the crazyness of many different shells was a lot confusing when I started writing scripts 2014-08-28 04:46:59 dalias, you are welcome to do it yourself 2014-08-28 04:47:02 One I liked for toys is mksh 2014-08-28 04:47:02 :) 2014-08-28 04:47:18 I have zero desire to hurt myself like that. 2014-08-28 04:47:33 If I don't like writing certain code, I will never write it at all 2014-08-28 04:47:37 dunno if it is secure, but it's a working-well pdksh rewrite 2014-08-28 04:47:47 And I hate writing POSIX shell code. 2014-08-28 04:49:56 Basically, the argument to only ever use POSIX shell code is like saying people must never use anything that is not available on a default system. 2014-08-28 04:50:07 Or, even worse, in all default systems. 2014-08-28 04:50:12 ash seems very fast, but lacks a lot for my habitudes in interactive mode 2014-08-28 04:50:18 That is not what portability means.; 2014-08-28 04:50:25 Who started that idea 2014-08-28 04:50:35 ^^' 2014-08-28 04:54:44 I'll take a nap 2014-08-28 04:54:50 ++ 2014-08-28 04:56:05 openfbtd, it depends on what you're doing 2014-08-28 04:56:34 if you're writing build scripts for an application that's supposed to be portable, there's a lot of value in not making the build-scripts non-portable 2014-08-28 04:56:48 it really sucks if you have to install some non-default tool just to use as part of building a package 2014-08-28 04:57:19 but if you're "writing an application in shell script", which sounds more like what you're doing 2014-08-28 04:57:30 then choose the shell language that works best for you 2014-08-28 04:57:53 goodbox, yes ash is an excellent shell 2014-08-28 04:58:14 it's fast, tiny, and seems to have few or no conformance bugs these days 2014-08-28 04:58:52 Obviously, build scripts should be POSIX... most of the time. 2014-08-28 04:58:57 i mildly prefer bash for interactive use, but there are a few things where i even prefer ash 2014-08-28 04:59:18 If they are complex enough to justify a non-standard scripting language being used — why not. 2014-08-28 04:59:29 i'm going look the man page in my bed 2014-08-28 04:59:56 For example, cjdns uses node.js both internally and for building. 2014-08-28 05:00:09 And that's ok, the build scripts for that thing are enormously complex 2014-08-28 05:00:56 Also if you're using the scripting language in runtime as well, there is no harm to also having it as a build dep 2014-08-28 05:01:01 You will still have to install it. 2014-08-28 05:01:18 well.. 2014-08-28 05:01:26 it rather sucks for anyone doing cross-compiling 2014-08-28 05:01:39 it means you need both the target and host version of the dep 2014-08-28 05:01:53 Not specifically. You still have to install a bunch of other crap. 2014-08-28 05:02:25 I would understand if we had a magical way of, say, building packages without having the build deps on the builders. 2014-08-28 05:03:02 But wether you are using clean environments or just building everything on a single system/chroot, you still need a bunch of stuff in there 2014-08-28 05:03:16 A scripting language is just another package 2014-08-28 05:04:22 I can understand POSIX shell scripts when they are small. Yes, it's silly to have zsh as a build dep just because you used a single zsh feature in a 50 line script 2014-08-28 05:04:44 But when we go thousands or even just hundreds of lines? Nope. Nope. Nope. 2014-08-28 05:09:39 well if your build script is that large, it probably means you failed to design it well :) 2014-08-28 05:30:21 Possible. It is also possible that your software is very complex and/or invasive, thus needing a complex build script 2014-08-28 15:59:07 is there someplace on internet (webgit or so) I could see the initscripts shipped with alpine? 2014-08-28 16:00:04 sorry 2014-08-28 16:00:25 http://git.alpinelinux.org/cgit/aports/log 2014-08-28 16:01:24 http://git.alpinelinux.org/cgit/aports/tree/main 2014-08-28 16:01:56 I guess http://git.alpinelinux.org/cgit/aports/tree/main/busybox-initscripts is what I'm looking for 2014-08-28 16:02:07 thought they'd be versioned separately 2014-08-28 16:07:07 CcxCZ: different packages will also include their own init scripts 2014-08-28 16:09:28 indeed. openrc and busybox are all that's interesting for me now I think. Alpine does not use netifrc (the classical openrc network setup), does it? 2014-08-28 19:23:55 Hi! I configured wpa_supplicant, I get a connection when I boot and 1min after the connection is dropped. I have a RTL8187 chipset, in iwevent I get a "wlan0 New Access Point/Cell address:Not-Associated". It is impossible to reconnect until I reboot. Maybe it's the driver? If so, how can i recompile the driver? 2014-08-28 22:01:48 o/ 2014-08-28 22:02:23 does anyone know a ash way to list functions in a script, like bash typeset -f ? 2014-08-29 00:44:42 i finally found a quick and dirty solution 2014-08-29 00:44:44 time grep -r -E "(^function\ )?[A-Z0-9a-z_]+\(\)" /etc/init.d/ | sed -e "s/()//g" -e "s/{//g" -e "s/}//g" 2014-08-29 00:45:42 took 0. 2014-08-29 00:45:46 0.02s * 2014-08-29 03:39:44 If I want use ccache, what I should do is to add export CC="ccache gcc" in /etc/abuild.conf ? 2014-08-29 06:04:40 wolas 2014-08-29 06:20:23 'morning 2014-08-29 06:24:48 morning 2014-08-29 06:49:39 saveis espaol? 2014-08-29 06:50:23 bueno digo 2014-08-29 06:50:52 por ke en linux mint 17 no me da error el virtual dj? 2014-08-29 06:51:06 me da error 2014-08-29 06:51:23 en virtualdj 2014-08-29 06:51:43 Hi errekerre, you might go on #linuxmint. 2014-08-29 06:51:53 vale 2014-08-29 07:11:36 errekerre: i dont speak spanish 2014-08-29 07:11:44 not sure we have many of them. 2014-08-29 07:11:54 and we prefer talk inhere is in english 2014-08-29 07:17:20 sip 2014-08-29 07:17:26 soy de espaa 2014-08-29 07:17:35 nu se ingles bro 2014-08-29 07:18:20 pero tengo linux 2014-08-29 07:19:30 ^^ 2014-08-29 07:20:51 usaste alpine-linux ? 2014-08-29 07:21:19 tengo linux-mint 17 2014-08-29 07:21:46 y entre en linuxmint 2014-08-29 07:21:53 pero no contesta nadie 2014-08-29 07:22:00 este canal no soporta linuxmint 2014-08-29 07:22:01 errekerre, esto es al canal the Alpine Linux distribution 2014-08-29 07:22:14 y de Alpine Linux Mail Client 2014-08-29 07:22:21 y no de Alpine Linux Mail Client* 2014-08-29 07:22:34 a ver el problema es que cuando estoy emitiendo por virtualdj se corta al de 4 canciones 2014-08-29 07:22:43 :') 2014-08-29 07:22:46 fijate ke meti el mIRC joni pack 2014-08-29 07:22:47 lol 2014-08-29 07:23:01 ke va genial y muchas cosas mas 2014-08-29 07:23:09 pero me falla el viertual dj 2014-08-29 07:23:26 lo demas va genial 2014-08-29 07:23:36 errekerre: /join #linuxmint 2014-08-29 07:23:38 wsupongo que sera igual 2014-08-29 07:24:04 ya Goodbox te ahorras con poner /j linuxmint 2014-08-29 07:24:07 ya entras 2014-08-29 07:24:18 vale muchas gracias 2014-08-29 07:24:32 de nada 2014-08-29 07:25:03 te ahorras letras oin # 2014-08-29 07:25:04 jijijijiji 2014-08-29 07:25:34 ._. 2014-08-29 07:29:43 bueno yo empezare a entrar por aqui 2014-08-29 07:29:51 a ver si aprendo ingles 2014-08-29 07:30:05 que ando muy torpe 2014-08-29 07:33:32 i only know: vamos a la playa 2014-08-29 07:35:06 wo wo wo 2014-08-29 07:35:14 que calienta el sol 2014-08-29 07:35:14 haha being french, i can partially understand the spanish vocabulary... 2014-08-29 07:35:21 y unos mojitos 2014-08-29 07:36:29 mojitos ? errekerre, no entendemos lo que quiere de nosotros. 2014-08-29 07:38:16 clandmeter dijo playa 2014-08-29 07:38:26 yo digo sombrilla y mojitos 2014-08-29 07:38:32 y tias buenas 2014-08-29 07:38:50 haha :) 2014-08-29 07:39:09 mojitos son ron blanco con yerva buena hielo picado y limon 2014-08-29 07:39:17 so why ddid he end up here? 2014-08-29 07:40:28 clandmeter: I'm not sure, a problem with linux mint 17 and a software called virtual dj 2014-08-29 07:41:13 :hum: 2014-08-29 07:47:52 does using utf-8 will impact performances on grepping files ? 2014-08-29 07:54:22 <^7heo> no. 2014-08-29 07:55:49 <^7heo> well, yes, because you are going to match more bytes... But it's not likely to be noticable. 2014-08-29 07:59:18 ok ! 2014-08-29 08:29:47 Hi ! 2014-08-29 08:30:50 <^7heo> hi 2014-08-29 08:32:05 Is there anything special I need to do to share my ports with the community ? 2014-08-29 08:32:14 Or a simple git repo is enough ? 2014-08-29 08:32:47 (eg: on crux, you can generate ports page like this: http://crux.z3bra.org/ports/) 2014-08-29 08:38:00 hi 2014-08-29 08:39:21 z3bra: for users who contribure frequent we tend to add a section to our git.a.o 2014-08-29 08:39:42 but there is nobody holding you back to setup your own or use someintg like github 2014-08-29 08:39:55 our aports repo is also on github for easy forking 2014-08-29 08:40:39 I plan on hosting them on my personnal git server 2014-08-29 08:40:53 but was just wondering about anything like arch's AUR or something 2014-08-29 08:41:01 but we dont have a special builder set for users contrib like aur 2014-08-29 08:41:09 okay, thanks ;) 2014-08-29 08:41:12 but 2014-08-29 08:41:18 it has been discussed before 2014-08-29 08:41:25 and it makes sence 2014-08-29 08:41:44 to remove crap from aports and move it to somwhere users can contrib 2014-08-29 08:41:45 yeah it could 2014-08-29 08:42:00 something like crux ports could be nice too 2014-08-29 08:42:46 users host their own ports 2014-08-29 08:43:07 and they are linked on the portdb at crux.nu 2014-08-29 08:44:16 hi z3bra ! 2014-08-29 08:44:35 and it could make the community more involved in package maintaining 2014-08-29 08:44:38 hi Goodbox 2014-08-29 08:48:32 z3bra: the thing is that somebody has to setup the infra 2014-08-29 08:48:48 its different then we use for our aports. 2014-08-29 08:49:06 yeha I understand 2014-08-29 08:49:41 how does aur/crux manage security? how can you add stuff? 2014-08-29 08:50:00 you don't "add" stuff 2014-08-29 08:50:29 s 2014-08-29 08:51:21 ? 2014-08-29 08:51:46 sorry surprise reboot ^^ 2014-08-29 08:51:57 you host you own ports 2014-08-29 08:52:05 and then mail the dev team to link 2014-08-29 08:52:33 to link your ports on the main site 2014-08-29 08:52:37 I have to go 2014-08-29 08:52:38 brb 2014-08-29 09:08:29 back 2014-08-29 09:12:15 z3bra: now i understand your question 2014-08-29 09:12:37 i was reading your first senctence incorrectly 2014-08-29 09:13:13 oh, sorry for not being clear enough then 2014-08-29 09:16:40 z3bra: you will also create your own builder? 2014-08-29 09:17:19 well, under crux we don't have "builder" 2014-08-29 09:17:22 but why not 2014-08-29 09:17:32 after all, it's just a for loop :P 2014-08-29 09:20:40 z3bra: does aur also provide packages? or is it just a repo? 2014-08-29 09:21:33 for archlinux ? 2014-08-29 09:21:51 it provides ports 2014-08-29 09:21:57 PKGBUILD + needed files 2014-08-29 09:22:09 and users have to compile them by themselves 2014-08-29 09:33:28 ok, i thought they provided the package repository aswell. 2014-08-29 09:34:19 nope 2014-08-29 09:34:32 never installed it. 2014-08-29 09:34:36 They build packages for main packages only 2014-08-29 09:37:54 so you prefer to share your own repo, or have it on git.a.o? 2014-08-29 09:37:57 <^7heo> git.a.o :D 2014-08-29 09:37:59 I don't know 2014-08-29 09:37:59 <^7heo> git.a.o :D :D 2014-08-29 09:37:59 I'll host them on my own repo anyway 2014-08-29 09:38:02 i think if you want to share it with alpine commuynity, git.a.o would make more sence 2014-08-29 09:38:11 <^7heo> sense 2014-08-29 09:38:11 <^7heo> but yeah 2014-08-29 09:38:19 It's up to what the maintainers want 2014-08-29 09:38:26 you are the maintainer :) 2014-08-29 09:38:26 well, it would make the whole repo a bit more "official" 2014-08-29 09:38:26 <^7heo> z3bra: I think that if you're having a repo on git.a.o, you're going to maintain that 2014-08-29 09:38:30 I'm the maintainer of my packages, not the distro 2014-08-29 09:38:31 <^7heo> there's no such thing as maintainers of "a distro" 2014-08-29 09:38:31 <^7heo> there are dictators, but that's political. 2014-08-29 09:38:31 else we would need to have a seperate page with users repo's 2014-08-29 09:38:47 <^7heo> the only thing is: 2014-08-29 09:38:48 Maybe a directory per user ? 2014-08-29 09:39:06 <^7heo> does your work "just work" everywhere, or not? 2014-08-29 09:39:26 <^7heo> if yes, you are a good dev/maintainer, and worth accepting packages from. 2014-08-29 09:39:26 <^7heo> if not, you're garbage, like most maintainers. 2014-08-29 09:39:26 eh 2014-08-29 09:39:26 define, "everywhere" 2014-08-29 09:39:26 <^7heo> everywhere as in: any linux based machine. 2014-08-29 09:39:26 <^7heo> you might add the "musl" requirement, since it's one of alpine anyway 2014-08-29 09:39:26 if you make quality pacakges that are always up to date, you should have commit access directly. 2014-08-29 09:39:26 <^7heo> and it's a sane one. 2014-08-29 09:39:26 no matter what the kernel, libc or userland is ? 2014-08-29 09:39:41 <^7heo> no matter what the (resonably recent) kernel, or userland is yes. 2014-08-29 09:39:49 <^7heo> and about the libc, I just defined what I was thinking :P 2014-08-29 09:40:11 ^^ 2014-08-29 09:40:17 <^7heo> because in the end 2014-08-29 09:40:30 <^7heo> 100% (rounding slightly up) of the distributions don't work. 2014-08-29 09:40:40 anyway, I'll mostly package softwares I'm not a dev of 2014-08-29 09:41:04 like libcaca, thttpd or goaccess 2014-08-29 09:41:04 <^7heo> yeah 2014-08-29 09:41:28 <^7heo> apk sucks, but it sucks 100% less than the alternatives. 2014-08-29 09:41:38 I like it 2014-08-29 09:41:43 it's not perfect, that's a fact 2014-08-29 09:41:44 <^7heo> it's not nearly sufficient. 2014-08-29 09:41:49 but it's pretty easy to use 2014-08-29 09:41:51 <^7heo> yeah 2014-08-29 09:41:54 <^7heo> that's the good part about it. 2014-08-29 09:41:58 <^7heo> it doesn't suck more than the rest 2014-08-29 09:41:58 I miss a few features tough 2014-08-29 09:42:05 <^7heo> but it is MUCH lighter. 2014-08-29 09:42:13 like listing the content of packages that are not installed 2014-08-29 09:42:27 thats by design 2014-08-29 09:42:29 <^7heo> so if you were to mesure the suckiness per byte, you would have a very low one on apk :P 2014-08-29 09:42:35 haha 2014-08-29 09:42:44 correct 2014-08-29 09:44:45 crux pkg system is fairly good to 2014-08-29 09:44:45 thats why we dont ship the db 2014-08-29 09:44:45 too much overhead for apk 2014-08-29 09:44:45 many program, each doing it's job 2014-08-29 09:44:45 <^7heo> z3bra: crux is source based, and at an end-user scope, that's insane, 2014-08-29 09:44:45 pkgadd, pkgrm, prtmk, ... 2014-08-29 09:44:45 That's a choice to make, obviously 2014-08-29 09:44:45 But I have been using it as my main distro since february, without problems 2014-08-29 09:44:45 z3bra: i think you can just setup your own repo and start hacking. When your face starts to become familiar we can add you to git.a.o. 2014-08-29 09:44:45 <^7heo> I was saying 2014-08-29 09:44:45 no prob ;à 2014-08-29 09:44:45 <^7heo> z3bra: crux is source based, and at an end-user scope, that's insane, 2014-08-29 09:44:45 <^7heo> I really want to write a package manager to distribute the work among all the participants from the original software developer to the end user. 2014-08-29 09:44:45 I'll setup a git repo, and give you the link then 2014-08-29 09:44:48 you mean, write THE package manager ? 2014-08-29 09:50:55 <^7heo> genau 2014-08-29 09:50:55 pkgmgr-final.tar.gz ? 2014-08-29 09:50:55 <^7heo> you got the idea. 2014-08-29 09:50:55 <^7heo> well 2014-08-29 09:50:56 you first have to finish that keyboard patch... ;) 2014-08-29 09:50:56 <^7heo> I'd like a solution that takes repositories, and turn them into installed software. 2014-08-29 09:50:56 <^7heo> clandmeter: it's done 2014-08-29 09:50:56 <^7heo> clandmeter: not pull-requested alright, but done. 2014-08-29 09:50:56 <^7heo> if you want the patches, I can send them. 2014-08-29 09:50:56 oh then i missed it 2014-08-29 09:50:56 repositories ? 2014-08-29 09:50:56 <^7heo> clandmeter: nah you didn't, I didn't send anything =/ 2014-08-29 09:50:56 no, i dont touch the kernel. 2014-08-29 09:50:56 else ncopa will kill me 2014-08-29 09:50:56 <^7heo> huhu 2014-08-29 09:50:56 <^7heo> clandmeter: due to the poor github interface. 2014-08-29 09:50:56 Oh, quick git question 2014-08-29 09:50:56 you dont need to use github, its there for people if its handy. 2014-08-29 09:50:56 <^7heo> clandmeter: I wanted to be the first one to do a pull request there :D 2014-08-29 09:50:56 ~/usr/aports is cloned from the alpine package repo 2014-08-29 09:50:56 <^7heo> clandmeter: but yeah, I don't have to, I know. 2014-08-29 09:50:56 <^7heo> Ubuntu sucks so hard... 2014-08-29 09:50:56 <^7heo> "The Desktop CD installer (ubiquity) currently does not provide support for setting up LVM or encrypted partitions." 2014-08-29 09:50:56 btw. why does alpine have separate archives for "control" and package contents? 2014-08-29 09:50:56 that seems like an overcomplicating 2014-08-29 09:50:56 *overcomplication 2014-08-29 09:50:56 I have ~/usr/aports/z3bra which is a freshly initialised git repo. But ~/usr/aports is already a git repo. So will I be able to manage the z3bra/ directory without problem ? 2014-08-29 09:50:56 and makes it harder to extract or create package archives 2014-08-29 09:50:56 <^7heo> z3bra: git usually goes up one directory after another until it finds a .git 2014-08-29 09:50:57 okay, so that will not mess with the "upper" repo ? 2014-08-29 09:50:57 <^7heo> z3bra: so if you are in ~/usr/aports/z3bra you should see only the z3bra git repo. 2014-08-29 09:50:57 <^7heo> not afaik. 2014-08-29 09:50:57 fine 2014-08-29 09:50:57 thanks 2014-08-29 09:50:57 <^7heo> but you might want to check submodules. 2014-08-29 09:50:57 That's what I'm doing 2014-08-29 09:50:57 I heard this word somewhere, but that's all :P 2014-08-29 09:51:29 why use submodules? 2014-08-29 09:53:37 <^7heo> in order to get used to them. 2014-08-29 09:53:37 <^7heo> otherwise, no real reason. 2014-08-29 11:56:18 And there you are: my personnal aports: http://git.z3bra.org/cgit.cgi/alpine/ 2014-08-29 12:00:36 \o/ 2014-08-29 12:01:23 how many personal aports repos exist? and what is the diff to the main ones (i have my own personal aports repo) 2014-08-29 12:02:19 i dunno 2014-08-29 12:02:35 i had one, but then i got commit acess 2014-08-29 12:03:01 i guess some have trees for business usage 2014-08-29 12:03:11 hidden from our eyes 2014-08-29 12:10:11 z3bra: is there a reason you want to keep them personal? your apkbuild's look clean and ready to be merged to main aports. 2014-08-29 12:32:34 well, you told me to maintain my own repo 2014-08-29 12:32:45 and when it's mature enough, you'll pass it to git.a.o 2014-08-29 12:33:58 Here is mine: https://github.com/jomat/aports 2014-08-29 12:45:36 clandmeter, but if you think my ports are ready to be merged, I'd be glad to do so :P 2014-08-29 13:22:00 Hey all, anyone around 2014-08-29 13:22:01 ? 2014-08-29 13:27:33 yep 2014-08-29 13:30:08 yes we are 2014-08-29 13:30:24 z3bra: yes im not saying you shouldnt :) 2014-08-29 13:30:49 I wonder if I could hit you up for a tip 2014-08-29 13:30:51 z3bra: but *if* you ever want it merged, your git tree should follow aports. 2014-08-29 13:31:00 trying to install alpine on a shitty pana toughbook. 2014-08-29 13:31:05 Has an intel 830m chip 2014-08-29 13:31:30 so have installed intel video driver package and checked dependencies 2014-08-29 13:31:42 but Xorg -configure errors with 'no device to configure' 2014-08-29 13:31:44 any clues? 2014-08-29 13:32:41 i dont run X on alpine. but i think ive done it a few times for intel in the past 2014-08-29 13:32:45 you have udev running? 2014-08-29 13:33:05 clandmeter, I did not know that 2014-08-29 13:33:11 I'll change my git tree then 2014-08-29 13:33:32 z3bra: you can clone our aports and add yours 2014-08-29 13:33:42 I did clone the aport tree 2014-08-29 13:34:00 but I added a "z3bra" directory aside from the "original" ones 2014-08-29 13:34:05 holding my packages 2014-08-29 13:34:17 But I'll change that quickly ;) 2014-08-29 13:34:40 generally we add packages to testing 2014-08-29 13:35:00 I noticed 2014-08-29 13:35:08 you can follow the same procedure. then when you want, you ask us to pull your tree 2014-08-29 13:35:15 so I only need a directory called "testing" 2014-08-29 13:35:27 do I need a github acount ? 2014-08-29 13:35:41 or you can pull from any repo ? 2014-08-29 13:36:18 we can pull from anywhere 2014-08-29 13:36:24 fine :) 2014-08-29 13:36:38 but if we pull, it needs to be from a aports copy 2014-08-29 13:37:06 we merge the changes in 2014-08-29 13:37:57 yeah sure 2014-08-29 13:38:14 I'll put my changes right in the aport repo then 2014-08-29 13:38:16 just git clone git.a.o/aports 2014-08-29 13:38:21 yup 2014-08-29 13:38:32 mkdir ~/aports/testing/mypkg 2014-08-29 13:38:48 WRONG 2014-08-29 13:38:49 or run newapkbuild inside testing 2014-08-29 13:39:04 cd ~/aports/testing && newapkbuild mypkg 2014-08-29 13:39:06 :P 2014-08-29 13:39:32 :) 2014-08-29 13:39:44 i stand corrected. its almost weekend 2014-08-29 13:39:57 ^^ 2014-08-29 13:44:48 so xorg log is saying that it can't relocate the *.so module 2014-08-29 13:44:52 that's the bit I'm stuck on 2014-08-29 13:44:55 is it a rights thing? 2014-08-29 13:45:22 Guest9405: what happends if you just start x? 2014-08-29 13:45:41 i remember i never used -configure to generate a config 2014-08-29 13:45:46 I get a black screen 2014-08-29 13:46:06 and what does the Xorg log say when you get a black screen? 2014-08-29 13:46:20 Guest9405: did you install and run udev? 2014-08-29 13:46:32 udev - is that a package? 2014-08-29 13:46:37 yes 2014-08-29 13:46:48 libudev or simply udev 2014-08-29 13:46:55 apk add udev 2014-08-29 13:47:37 apk infp | grep udev 2014-08-29 13:47:42 libudev 2014-08-29 13:47:44 udev 2014-08-29 13:47:47 libgudev 2014-08-29 13:47:57 info* not infp 2014-08-29 13:48:30 so I've run X 2014-08-29 13:48:46 a load of lines scrolled off befor eI could read then, but there was alot of 'initialising' going on 2014-08-29 13:48:59 no HDD activity, black screen 2014-08-29 13:49:00 apk add udev 2014-08-29 13:49:00 /etc/init.d/udev start && /etc/init.d/udev-postmount start 2014-08-29 13:49:00 rc-update add udev sysinit 2014-08-29 13:49:01 rc-update add udev-postmount default 2014-08-29 13:49:15 right, I will try that. brb 2014-08-29 13:49:59 Guest9405: dont forget apk add xf86-input-evdev 2014-08-29 14:39:52 so I've still got a black screen (I had to pop out) - how do I drop out of that. Do I have to reboot it? 2014-08-29 14:40:10 before I can try what clandmeter has suggested 2014-08-29 14:43:50 apologies for noob. I've not used busybox/alpine before. It's great it's so cut down, but I've got used to linux in easy mode :P 2014-08-29 14:44:14 SysRq-E 2014-08-29 14:45:18 a reboot takes ~ 3 minutes. It's tortuous on this box :P 2014-08-29 14:45:33 <^7heo_> Guest27896: with alpine?! 2014-08-29 14:46:47 yep 2014-08-29 14:46:56 it's a PIII 800Mhz 2014-08-29 14:47:01 back to the 90s! 2014-08-29 14:47:28 once it posts, alpine actually starts quite quickly. 2014-08-29 14:47:55 it dicks around for ages before the bootloader is even sniffed at 2014-08-29 14:49:09 "/etc/init.d/udev start && /etc/init.d/udev-postmount start@ - already started 2014-08-29 14:49:34 rc-update - already installed in sysinit 2014-08-29 14:49:58 udev-postmount in runlevel default 2014-08-29 14:51:06 input-evdev is installed 2014-08-29 14:53:22 Guest27896: what's in your ~/.xinitrc? 2014-08-29 14:58:00 .xinitrc doesn't exist 2014-08-29 14:58:17 in my home dir 2014-08-29 14:58:19 sec, let me search 2014-08-29 14:58:22 nope 2014-08-29 15:00:38 scanning the Xorg log, it seems fbdev module is failing, so I'm going to try and grab that 2014-08-29 15:04:02 ok, so it's correctly identified my chipset 2014-08-29 15:04:27 I wonder if it's trying to set a resolution that this p.o.s. can't display 2014-08-29 15:04:51 I also got a blank screen until i managed to start my window manager :-) 2014-08-29 15:05:51 [ 25.210] (EE) Failed to load module "fbdev" (module does not exist, 0) 2014-08-29 15:05:58 That's ok 2014-08-29 15:06:49 I've now installed fbdev.so 2014-08-29 15:07:06 ok, so the res is capable of 800x600, which I can see is being selected in the log file 2014-08-29 15:08:33 how do you start the window manager 2014-08-29 15:09:08 i. e.: exec /usr/bin/awesome 2014-08-29 15:09:31 sec, need to reboot the sodding thing, as I've got black screen 2014-08-29 15:09:35 ACTION waits 2014-08-29 15:09:44 Guest27896: Kill it with sysrq-e 2014-08-29 15:10:02 that kills everything, but is still faster then rebooting 2014-08-29 15:10:28 just type that into console without any ctrl keys? 2014-08-29 15:10:57 Guest27896: If you don't have a .xinitrc, /etc/X11/xinit/xinitrc will try to startxfce4 2014-08-29 15:11:14 so another approach could be to install xfce4 for testing purposes 2014-08-29 15:11:38 I have put the xfce4 package on there 2014-08-29 15:11:49 Na, it's some key combo... on the german keyboard layout something like ctrl-alt-print-e 2014-08-29 15:12:11 ah right 2014-08-29 15:12:19 here we go, it's strapping 2014-08-29 15:13:13 so if startxfce4 is in your path it should be ok 2014-08-29 15:13:51 xorg.conf.d filder exists in /etc/X11 2014-08-29 15:13:54 folder* 2014-08-29 15:14:04 there is a file called 20-modules.conf 2014-08-29 15:14:33 running ./xinitrc 2014-08-29 15:14:38 blimey, it only works 2014-08-29 15:15:50 holy cow, the touchscreen works as well 2014-08-29 15:17:42 it needs recalibrating it seems. 2014-08-29 15:17:47 but: awesome, thank you 2014-08-29 15:17:55 that's the way to start it, not Xorg 2014-08-29 15:23:09 nice \o/ 2014-08-29 15:24:40 open RC is an unsual envt for me 2014-08-29 15:24:49 different words for reboot and shutdown etc. 2014-08-29 15:40:38 hm. 2014-08-29 15:40:52 there isn't an rc file - should I create one called .ash_rc? 2014-08-29 15:41:13 so that I can tell it to put /etc/X11/xinit in $PATH every log in 2014-08-29 15:49:55 actually 2014-08-29 15:49:58 startx works 2014-08-29 15:50:19 <^7heo_> startx sucks 2014-08-29 15:50:30 and I've now created a .xinitrc with exec startzfce4 and setxkbmap to gb 2014-08-29 15:50:32 job is a good one 2014-08-29 15:50:35 how come it sucks? 2014-08-29 15:50:44 startxfce4* 2014-08-29 15:50:52 seems to work 2014-08-29 15:51:06 <^7heo_> it sucks because it cannot take a user as a parameter 2014-08-29 15:51:24 <^7heo_> so you cannot start it in the user you want without logging the user in. 2014-08-29 15:51:31 ah I see 2014-08-29 15:51:36 I think we'll only be runnign as root tbh 2014-08-29 15:51:37 <^7heo_> So in the end 2014-08-29 15:51:43 <^7heo_> if you want startx to start in !root 2014-08-29 15:51:49 but I'm willing to heaer another way 2014-08-29 15:51:58 hear* 2014-08-29 15:52:41 the other thought I had was to create a .bashrc or .bash_profile or .ashrc whichever it is to add the /etc/X11/xinit to $PATH 2014-08-29 15:52:43 <^7heo_> Well, I didn't find yet. 2014-08-29 15:52:45 <^7heo_> unless you use ksm 2014-08-29 15:52:47 <^7heo_> kms even 2014-08-29 15:52:59 and then can start using xinit - or is that the same as startx? 2014-08-29 15:53:40 <^7heo_> the point is that you startX without startx, but with your x login manager 2014-08-29 15:54:14 hm. 2014-08-29 15:54:24 ok, I'll look into that. 2014-08-29 15:54:31 thank you all for your help. 2014-08-29 15:54:39 at least I have a working gui envt 2014-08-29 15:54:48 <^7heo_> and therefore, the rest of the X applications will be started as your logged user 2014-08-29 15:55:11 <^7heo_> but I'm not a big fan of having X and the GUI login manager running as root, personally. 2014-08-29 15:55:12 yeh, I see. so it's not ideal to start as root realy 2014-08-29 15:55:18 really* 2014-08-29 15:55:20 ^^ 2014-08-29 15:55:24 <^7heo_> You can start X as !root, but again, it's using kms 2014-08-29 15:55:29 <^7heo_> so it's a different setup. 2014-08-29 15:55:52 <^7heo_> https://en.wikipedia.org/wiki/Mode_setting 2014-08-29 15:56:29 I think I see what you mean - I need to use UMS? 2014-08-29 15:56:53 or am I confused as to the meaning 2014-08-29 15:57:41 <^7heo_> not necessarily 2014-08-29 15:57:49 <^7heo_> https://wiki.ubuntu.com/X/Rootless 2014-08-29 15:57:57 <^7heo_> Sorry to link to noobuntu 2014-08-29 15:58:06 <^7heo_> but it's not a widly documented topic yet 2014-08-29 15:58:19 I have only used ubuntu and debian up until now. 2014-08-29 15:58:20 <^7heo_> http://lwn.net/Articles/341033/ 2014-08-29 15:58:31 Linux in easy mode - is why I''m struggling a little here with alpine :P 2014-08-29 15:58:44 <^7heo_> Debian sucks, and Ubuntu sucks even more 2014-08-29 15:58:49 hehe 2014-08-29 15:58:58 <^7heo_> Don't get me wrong 2014-08-29 15:59:01 <^7heo_> I'm using debian right now 2014-08-29 15:59:04 everyone's a critic. 2014-08-29 15:59:16 2014-08-29 15:59:20 <^7heo_> mostly because there's no support for my hardware in anything earlier than 3.13.0.35 2014-08-29 15:59:40 <^7heo_> and that Alpine isn't mature enough to support the sound/video of my CPU 2014-08-29 15:59:51 <^7heo_> it doesn't even support the keyboard yet... 2014-08-29 16:00:06 ^7heo - this is interesting. However, I've got to go for dinner right now. Thanks for the hints, I will check 'em out 2014-08-29 16:00:15 take it easy o7 2014-08-29 16:01:04 <^7heo_> Guest27896: dinner at 18? 2014-08-29 16:01:10 <^7heo_> alright, guten appetite. 2014-08-29 16:01:28 <^7heo_> s/te.$/t./ 2014-08-29 16:18:14 I have been using debian and I'm looking for a more up-to-date, and secure linux distribution for general usage. btw. I can't find an Emacs package in your repository ...? do you think Alpine is right for me? 2014-08-29 16:18:45 <^7heo_> gr8: would you maintain emacs? 2014-08-29 16:19:12 hehe. that means you don't have many people 2014-08-29 16:19:53 I have no idea of maintaining software, so ... 2014-08-29 16:25:14 <^7heo_> it's easy 2014-08-29 16:25:24 <^7heo_> and yes, we're like debian 2014-08-29 16:25:31 <^7heo_> lacking skilled, committed people. 2014-08-29 16:25:46 <^7heo_> or like ubuntu, also. 2014-08-29 16:26:03 <^7heo_> in ubuntu it's much more visible because they have a lot of unskilled, uncommitted people. 2014-08-29 16:28:32 Debian is lacking skilled, committed people? I always had a different impression, but that's interesting 2014-08-29 16:28:41 <^7heo_> yeah 2014-08-29 16:28:53 <^7heo_> they have a lot of skilled, or committed, but not many of both 2014-08-29 16:29:33 <^7heo_> the implied conjuction was an and, not an or. 2014-08-29 16:30:17 implied conjunction. so you're a CS person! xD 2014-08-29 16:30:41 conjunction always means 'and' 2014-08-29 16:31:22 <^7heo_> Indeed. I couldn't be an engineer, engineers are required to give up both mental sanity and professional conscience. 2014-08-29 16:31:44 <^7heo_> No 2014-08-29 16:31:46 <^7heo_> "Or 2014-08-29 16:31:48 <^7heo_> damn 2014-08-29 16:31:54 <^7heo_> "Or" also is a conjuction. 2014-08-29 16:31:58 <^7heo_> https://en.wikipedia.org/wiki/Conjunction_%28grammar%29 2014-08-29 16:32:08 yes you're right. just read that article too :D 2014-08-29 16:32:17 <^7heo_> :D 2014-08-29 16:33:38 unfortunately, I'm not really skilled yet, only commited 2014-08-29 16:33:47 <^7heo_> That's a good start 2014-08-29 16:33:55 <^7heo_> that's how you get skilled with time. 2014-08-29 16:34:13 <^7heo_> Unfortunately, I think that having only commited people contributing is harmful. 2014-08-29 16:34:26 <^7heo_> (or, "uninterestingly enough" rather) 2014-08-29 16:35:26 <^7heo_> Because I have no authority here. 2014-08-29 16:35:57 a lack of skilled people is always harmful to a project, right? 2014-08-29 16:36:10 <^7heo_> yeah 2014-08-29 16:36:15 <^7heo_> especially when those people are in charge of something 2014-08-29 16:36:16 <^7heo_> anyway 2014-08-29 16:36:18 <^7heo_> I gtg 2014-08-29 16:36:19 <^7heo_> o/ 2014-08-29 16:36:22 <^7heo_> take good care 2014-08-29 20:03:04 i have a pre-compiled binary file that i'm trying to execute (./binaryfile), but i get this error message: "-ash ./binaryfile: not found" 2014-08-29 20:03:26 i'm inside the same directory that the file is in, and the permissions are correct 2014-08-29 20:36:16 oM 2014-08-29 20:36:18 o/ * 2014-08-29 20:36:35 catcow: does he mounted from another partition/media ? 2014-08-29 20:37:27 no, i'm just in my home directory 2014-08-29 20:37:35 catcow: Is it compiled with musl? 2014-08-29 20:37:44 ok, what's elf type ? 2014-08-29 20:38:14 it is not compiled with musl 2014-08-29 20:38:50 i tried compiling the source code but i couldn't get it to work, so i'm trying pre-compiled binaries (hoping that it would be easier!) 2014-08-29 20:39:01 Goodbox, i do not know what an elf type is 2014-08-29 20:39:28 ACTION googles... 2014-08-29 20:40:05 catcow: it's a executable format which support static or dynamic loadables libraries 2014-08-29 20:41:00 by typing file you should get the format 2014-08-29 20:42:09 'file' is a command? 2014-08-29 20:42:20 yes 2014-08-29 20:42:59 it's a classic unix tool attemptin to detail files types and contents 2014-08-29 20:43:02 +g 2014-08-29 20:43:30 nice, just installed it. i have to type out the input because i'm working in a VM 2014-08-29 20:45:09 "ELF 64-bit LSB executable, x86-64, version 1 (GNU/Linux), dynamically linked (uses shared libs), for GNU/Linux 2.6.24, BuildID[sha1]=...., not stripped 2014-08-29 20:45:17 " 2014-08-29 20:45:19 then there is few reasons why it could work 2014-08-29 20:45:31 can you paste us the output of uname -a command ? 2014-08-29 20:45:49 couldn't 2014-08-29 20:46:26 "Linux wallet 3.14.17-0-grsec #1-Alpine SMP Tue Aug 26 08:50:11 GMT 2014 x86_64 Linux" 2014-08-29 20:48:42 thanks, one causes, is the absence of required shared libs 2014-08-29 20:48:56 as you see, he isn't static 2014-08-29 20:52:41 yea 2014-08-29 20:52:55 is there a way to find out what is required? 2014-08-29 20:52:55 catcow: readelf -l binaryfile|grep interpreter 2014-08-29 20:53:25 i have installed the dependencies that were listed on the program's website 2014-08-29 20:53:33 catcow: ldd 2014-08-29 20:53:34 catcow: readelf is in binutils 2014-08-29 20:53:53 ldd will yield: 2014-08-29 20:53:54 ldd: more: Not a valid dynamic program 2014-08-29 20:54:14 ehh... s/more/binaryfile/ 2014-08-29 20:55:23 jomat: "[Requesting program interpreter: /lib64/ld-linux-x86-64.so.2]" 2014-08-29 20:55:53 Goodbox: "ldd: simplewallet: Not a valid dynamic program" 2014-08-29 20:56:14 just like you said, jomat 2014-08-29 20:57:39 ok, does another problem could be the kernel version compatibility ? 2014-08-29 20:58:20 If it's compiled for musl it should be: Requesting program interpreter: /lib/ld-musl-x86_64.so.1 2014-08-29 20:59:00 hm, i installed alpine with the default settings 2014-08-29 21:02:13 can you tell us which software you try to use? 2014-08-29 21:03:10 https://github.com/monero-project/bitmonero 2014-08-29 21:03:35 they have an irc channel here on freenode, i will try asking there 2014-08-29 21:04:41 musl compatibility can be roughly checked here: http://wiki.musl-libc.org/wiki/Pkgsrc_results 2014-08-29 21:05:43 thx jomat, I was expecting that for one day ! 2014-08-29 21:06:58 http://www.musl-libc.org/faq.html says: "At present, some glibc-linked shared libraries can be loaded with musl, but all but the simplest glibc-linked applications will fail if musl is dropped-in in place of /lib/ld-linux.so.2." 2014-08-29 21:08:23 many gnu and others programs have glibc asserts :/ 2014-08-29 21:16:31 so i have to compile the source so that it uses musl 2014-08-29 21:17:13 yes you can try it! 2014-08-29 21:18:58 sometimes, just patching the makefile according to musl needs can works, however there's a chance for glibc assert presence in the project. 2014-08-29 23:21:24 does anyone know how to change the default ulimits on alpine? 2014-08-29 23:21:41 there isn't an /etc/limits.conf or /etc/security... so im not sure where to begin to look 2014-08-29 23:38:24 hmm :) busybox seems do not support ulimit 2014-08-29 23:39:29 maybe installing pam ? 2014-08-29 23:42:18 ah linux-pam bring /etc/sercurity folder 2014-08-29 23:42:48 i think just a reloging is needed after changes 2014-08-30 00:51:43 ok, ill try it out 2014-08-30 00:51:48 i ll need pam for i3lock anyways 2014-08-30 00:56:57 ok 2014-08-30 00:58:05 pam seem to be nice and, it's really helpful to set limits or authenticate trough the web 2014-08-30 01:03:58 otherwise on my personnal system I don't use lockscreens and not a forkbombs subject ^^. I guess open-rc script and login scripts are sufficient to define ressources and limits with simple tools. 2014-08-31 00:28:05 hello everyone. I'm looking for a good S.O that runs on RAM to build a plex media server and came across alpine linux, can anyone tell me if Alpine is a good S.O for my needs? 2014-08-31 03:07:34 morning 2014-08-31 08:40:28 what is a S.O? 2014-08-31 08:41:22 dunno anymore 2014-08-31 09:52:18 could be "operating system" in romance languages 2014-08-31 09:53:03 ehehe :) 2014-08-31 09:53:25 Systeme Operationelle? 2014-08-31 09:58:37 sort of. http://it.wikipedia.org/wiki/Sistema_operativo http://es.wikipedia.org/wiki/Sistema_operativo http://pt.wikipedia.org/wiki/Sistema_operativo http://ro.wikipedia.org/wiki/Sistem_de_operare etc. 2014-08-31 10:02:45 seems legit, the IP is from protugal. 2014-08-31 10:08:26 haha 2014-08-31 10:11:19 système d'exploitation in french, but nobody write S.E 2014-08-31 10:11:46 Plex devs don't know of musl (or alpine) 2014-08-31 10:12:52 A lot of people don't know of musl 2014-08-31 10:13:06 Including a lot of devs 2014-08-31 10:13:32 yeah, best things are small 2014-08-31 10:15:09 Like breasts 2014-08-31 10:15:18 I will show myself out 2014-08-31 10:15:44 bwaha 2014-08-31 10:19:33 what are the chances of a random modern binary (say from a commercial product) working with musl? 2014-08-31 10:20:30 I dunno, except qt clients it should be buildable trough musl-gcc in theory :s 2014-08-31 10:20:48 I turned around some gcc-libc and glib alternatives. Musl bite all others in dependecy and pureness 2014-08-31 10:20:58 Most complex stuff I tried doesn't build like that. 2014-08-31 10:21:21 i'm asking about an already built binary, let's say lsb glibc 2.2.5 to be precise 2014-08-31 10:21:42 or is musl not binary compatible with glibc at all? 2014-08-31 10:21:58 openfbtd: using musl-gcc ? 2014-08-31 10:22:22 Yep 2014-08-31 10:23:34 ok, it's a pity 2014-08-31 10:24:14 brb, my kitchen's fuse jumped 2014-08-31 10:34:20 anyone can precise me, about what's the main compatibility problem between musl and glib ? I experimented how far glib is bloated for building static executables, but they're both designed for linux calls isn't ? 2014-08-31 10:37:42 so as not to cause confusion, note that glib is a completely separate project, you're talking about glibc. 2014-08-31 10:39:09 Ok, sorry for my misunderstandings, then the gcc's libc, glibc and glib are different things 2014-08-31 10:39:24 there's also eglibc and eglib :) 2014-08-31 10:39:47 gcc doesn't have a libc, it just has libgcc 2014-08-31 10:40:08 thanks for clarification 2014-08-31 10:59:01 there's a musl compatibility wiki page and a faq for exactly these kind of questions 2014-08-31 10:59:16 not this one, the one on musl-something. 2014-08-31 10:59:59 http://wiki.musl-libc.org/wiki/Pkgsrc_results http://www.musl-libc.org/faq.html 2014-08-31 11:35:45 f15RyWpgGTO6: thanks 2014-08-31 11:57:42 thanks you ! sorry for dumbs again 2014-08-31 14:49:49 Hi 2014-08-31 17:57:50 Hi 2014-08-31 17:58:21 I can't find the doc about linux-vanilla kernel installatiions steps 2014-08-31 17:58:49 does somebody knows howto install this package ? 2014-08-31 18:06:46 I use alpinelinux on a usb key so /lib/firmware is ro and I can't find howto remount /.modloop rw 2014-08-31 18:53:40 Forth: You just have to install the linux-vanilla package 2014-08-31 18:53:54 http://git.alpinelinux.org/cgit/aports/tree/main/linux-vanilla 2014-08-31 18:54:43 what is modloop ? 2014-08-31 18:55:21 mount -o remount,rw doesn't works ? 2014-08-31 18:55:52 Goodbox, nop 2014-08-31 18:56:25 red:~# apk add linux-vanilla 2014-08-31 18:56:26 (1/4) Installing lddtree (1.25-r1) 2014-08-31 18:56:26 (2/4) Installing mkinitfs (2.6.0-r5) 2014-08-31 18:56:26 (3/4) Installing linux-firmware (20140507-r0) 2014-08-31 18:56:26 ERROR: Failed to extract lib/firmware/LICENCE.ti-connectivity: Read-only file system 2014-08-31 18:56:26 ERROR: linux-firmware-20140507-r0: Read-only file system 2014-08-31 18:56:28 (4/4) Installing linux-vanilla (3.14.17-r0) 2014-08-31 18:56:30 ERROR: Failed to extract lib/modules/3.14.17/modules.symbols: No such file or directory 2014-08-31 18:56:32 wget: short write 2014-08-31 18:56:34 ERROR: linux-vanilla-3.14.17-r0: No such file or directory 2014-08-31 18:56:36 Executing busybox-1.22.1-r9.trigger 2014-08-31 18:56:36 dmesg doesn't say why ? 2014-08-31 18:56:38 2 errors; 240 MiB in 93 packages 2014-08-31 18:57:08 nope 2014-08-31 18:57:43 I never used alpine on usb key, except for install :/ 2014-08-31 18:58:19 I'm trying to use cgminer 2014-08-31 18:58:24 but it sergfault 2014-08-31 18:58:34 and I wanted to try linux-vanilla 2014-08-31 18:59:04 If you are using it like a live, you'll must kexec the new kernel 2014-08-31 19:02:16 does linux-firmware* appears out the command mount ? 2014-08-31 19:03:07 red:~# ls -alh /lib/ 2014-08-31 19:03:08 total 3212 2014-08-31 19:03:08 drwxr-xr-x 6 root root 500 Aug 31 20:28 . 2014-08-31 19:03:08 drwxr-xr-x 19 root root 400 Aug 31 20:56 .. 2014-08-31 19:03:08 drwxr-xr-x 3 root root 60 Aug 31 20:28 apk 2014-08-31 19:03:08 lrwxrwxrwx 1 root root 21 Aug 31 20:28 firmware -> /lib/modules/firmware 2014-08-31 19:03:17 lrwxrwxrwx 1 root root 17 Aug 31 20:28 modules -> /.modloop/modules 2014-08-31 19:03:48 isn't in the output of mount ? 2014-08-31 19:04:05 linux-firmware ? 2014-08-31 19:04:24 /lib/modules or /.modloop 2014-08-31 19:05:01 isn't just your rootfs ? 2014-08-31 19:05:41 /dev/loop0 on /.modloop type squashfs (ro,relatime) 2014-08-31 19:05:55 okay 2014-08-31 19:05:57 what about / 2014-08-31 19:06:05 rw 2014-08-31 19:06:10 hmm 2014-08-31 19:06:43 I'm damn new on alpine, i have no ideas, sorry... 2014-08-31 19:06:49 I switch ro to rw in /etc/runlevels/sysinit/modloop 2014-08-31 19:06:51 Maybe considering an installation ? 2014-08-31 19:06:55 didn't help 2014-08-31 19:07:09 I can't, the sata port is nok 2014-08-31 19:07:18 shit :s 2014-08-31 19:07:55 I should have done my bga reballing homework 2014-08-31 19:09:05 well, my main problem is cgminer segaulting 2014-08-31 19:09:08 :D 2014-08-31 19:09:25 segfault error 6 2014-08-31 19:09:38 I've tried to trace it, no luck 2014-08-31 19:09:46 s/trace/strace/ 2014-08-31 19:10:06 yarf, probably due of an glibc assert 2014-08-31 19:10:21 lsattr /.modloop ? 2014-08-31 19:10:23 :'( 2014-08-31 19:13:50 Probably the better way would to adapt cgminer for musl, but for my part, i'm unable ! :p 2014-08-31 19:14:46 cgminer is compiling well 2014-08-31 19:15:04 I bet there is something with pax/grsecurity 2014-08-31 19:15:38 yeap, he might pee somewhere, maybe trying to pass valgrind will show more ? 2014-08-31 19:17:41 indeed 2014-08-31 19:18:07 I'm a bit short on time unfornutaly 2014-08-31 19:18:47 spend this afternoon ducktaping a garbaged laptop to boot alpine and running cgminer 2014-08-31 19:19:06 i'm shure you'll find someone here who can help you 2014-08-31 19:19:14 haha :) 2014-08-31 19:21:01 you're crazy dude 2014-08-31 19:21:49 that's just for fun 2014-08-31 19:22:08 on the job side I've build 2014-08-31 19:22:12 hmm wait, I think i have ane idea 2014-08-31 19:22:48 a xen rig of alpine usbstick without disk and an iscsi san (freenas) 2014-08-31 19:22:49 if my memory works, there's somewhere in gentoo profiles 2014-08-31 19:22:53 a musl one 2014-08-31 19:22:59 with packages 2014-08-31 19:23:12 i'm going to check for cgminer 2014-08-31 19:24:12 thx, I'm not familiar with gentoo 2014-08-31 19:29:17 idem, i tried to familiarize with emerge, but's a little bit unaprehendable 2014-08-31 19:30:03 http://packages.gentoo.org/package/net-misc/cgminer visibly, no patches for hardened/musl 2014-08-31 19:31:37 i'm searching how to search in bugs relative to pax and grsec 2014-08-31 19:32:13 however i'm really slow at this time 2014-08-31 19:34:03 Goodbox, don"t go further 2014-08-31 19:34:21 bwarf "Gentoo's Bugzilla needs a legitimate login and password to continue." 2014-08-31 19:34:22 I've forgotten that I don't use the official source but a fork 2014-08-31 19:34:26 only in two searchs x) 2014-08-31 19:34:42 okay 2014-08-31 19:34:57 hopping the official works :) 2014-08-31 19:35:59 but i'm incertain, i dont know even if scrypt works on alpine 2014-08-31 19:37:46 if it helps you to investigate https://bugs.gentoo.org/buglist.cgi?quicksearch=grsec 2014-08-31 19:39:07 or it just an incompatibility with musl http://wiki.musl-libc.org/wiki/Functional_differences_from_glibc 2014-08-31 19:39:13 +is 2014-08-31 19:44:30 time to bed, bye ! 2014-08-31 19:50:07 bye